c9xX9mhc

1~somlus Members: I promise to out-stupid you with stupid questions. This
2        is a chat room for people who have NEVER coded but want to learn
3        hoon, and the saints who flutter in to show them patience.
4 
5        In a way I'm hoping that this uber basic lens of dialogue will
6        ultimately help create a set of hoon education documents for the
7        very very early coder.
8 
9        If people get mad at the 'tards' i'll change it.
10 
11        codiots was the competing name...
12 
13~risruc that's for setting this up ~somlus. I was thinking of the need
14        for the same myself.
15 
16        cheers!
17 
18~somlus cool! let's hope to find a code saint on urbit help who can
19        answer questions here when we are posting. it seems like you
20        actually have a decent base in this? maybe we can come up with
21        a few very easy generators like the list one but also others
22        which are very simple and then learn the basics writing those
23        together.
24 
25~tirlyn I can help a bit
26 
27        I'm not always on but I read the scroll back.
28 
29~somlus thank you very much tirlyn!
30 
31~ribreg Visual Studio Code - Installed. home folder - mounted.
32 
33        (this is a hoontard comet of somlus)
34 
35        Hoontard questions, this is confirming the likely obvious.
36        Please tell me if these are correct: Is a 'natural number' an
37        integer? Is an unsigned number basically a positive number?
38 
39~somlus hmm the hoontard comet is slow. switching back to somlus. is
40        a child sort of a highly variable concept, could be a few
41        characters, could be a few lines, could be a gigantic swath of
42        code, and only by knowing what the gate does can you have a
43        reasonable expectation as to what the child is going to look
44        like / consist of?
45 
46~risruc yes, natural number means positive whole numbers.
47 
48        unsigned means without sign
49 
50        but 'positive' is close enough.
51 
52        signed integers in Hoon take twice as many bits to define, since
53        every other one is negative.
54 
55        it's worth looking at how negative numbers cast to ud
56 
57~tirlyn a natural number is strictly 0 or greater whole numbers only
58 
59        in urbit everything is a unsigned int at the bottom, any other
60        representation is an aura on top
61 
62        basically a type that says "this is how this int should be
63        interpreted
64 
65~risruc `@ud`-1
66 
67        `@ud`-2
68 
69        `@ud`-2
70 
71~tirlyn ahh sorry risruc didn't mean to repeat what you said.
72 
73~risruc nope, appreciated. the chat was a bit slow for a minute.
74 
75~somlus This is super helpful. Can you clarify "is a unsigned int at
76        the bottom, any other representation is an aura on top" -
77        specifically "at the bottom" and "on top". And is an "aura"
78        just a way of representing an integer in one of many different
79        formats for integers?
80 
81        or rather, a system of interpreting an input which could be
82        one value in one representation but would be another value in
83        another system? I.e. 110 means very different stuff in binary vs
84        decimal vs hex?
85 
86~risruc yes, that's correct.
87 
88        "at the bottom" -- in essence
89 
90        "on top" -- as interpreted by the compiler.
91 
92        which aren't technical hoon terms, by the way.
93 
94        but yes, you got the point. many basic types are all 'actually'
95        integers
96 
97        the compiler also infers (or is told) their aura, which it uses
98        to render them in more meaningful ways.
99 
100        for example, ship names are @p.
101 
102        ? ~somlus-savlev
103 
104        but if you cast them to an unsigned decimal..
105 
106        `@ud`~somlus-savlev
107 
108~somlus Awesome!
109 
110~risruc so really your address is 1838679296
111 
112~tirlyn the example of unsigned integer versus signed gets at the
113        relationship. signed integers use a binary representation called
114        twos complent, this means if you cast an signed integer to a
115        unsigned integer there are many cases where you would get a
116        different value given the same bits
117 
118        you correctly point out that the representation of binary and
119        hex express encode no extra data so you can freely cast between
120        them
121 
122        for a really concrete example what happens if you cast an signed
123        in (like --3, that is positive 3) to binary and compare it to
124        the unsigned equivalent?
125 
126        another interesting example is floating point numbers, if you
127        cast one to a unsigned int you will find that it is not the
128        number you may have expected, because the binary representation
129        of a float is very different then that of an int.
130 
131~risruc that's just not as easy to remember as ~somlus-savlev
132 
133~somlus can you clarify "casting" a signed integer to an unsigned
134        integer?
135 
136~risruc to cast to a certain aura, use `@aura`atom
137 
138~ronreg ~somlus: so, hoon compiles to nock, and part of the nock
139        specification is that the most fundamental units of computation
140        are natural numbers and ordered pairs of natural numbers
141 
142        https://urbit.org/docs/tutorials/nock/definition/
143 
144~somlus i did recall reading that ~ronreg, and got bogged down in a
145        simple program as to how a subtraction is done in nock. That's
146        another question on this forum for another day, but I think
147        someday I need to understand that at its core to get all this.
148 
149        So then how does the computer know that one representation
150        is a negative number but another is a positive number (of a
151        completely different value), based on the binary representation
152        of those two numbers?
153 
154~risruc in my opinion, you don't.
155 
156~somlus good to hear.
157 
158~risruc it's useful to understand as much as ~ronreg said... i.e., that
159        everything compiles down to nock operations on integers and
160        pairs of integers.
161 
162        but you do not need to learn the nock codes.
163 
164        i've tried studying them a bit, but don't have a grasp.
165 
166        and that's at least loosely within my area of interest
167 
168~somlus got it. note to self: create nocktard forum in 2021.
169 
170~risruc hahaha
171 
172~somlus There was one more question up there: is a child sort of a
173        highly variable concept, could be a few characters, could be
174        a few lines, could be a gigantic swath of code, and only by
175        knowing what the gate does can you have a reasonable expectation
176        as to what the child is going to look like / consist of?
177 
178~risruc a child is always one hoon.
179 
180        however....
181 
182        ;)
183 
184~somlus now yer just talkin' santa claus
185 
186~risruc you're half way decent with excel, i take it.
187 
188~somlus yes
189 
190        to be literal.
191 
192~risruc OK. good. it's easy to analogize.
193 
194        in excel, every function takes a fixed number of arguments.
195 
196        (most*, you know what i mean).
197 
198~somlus yes definitely
199 
200~risruc if you want to return X at max Y, you might do something like:
201 
202~somlus and there is a very distinct structure with brackets, commas,
203        parentheses
204 
205~risruc =max(b1:b1000) to get the max. maybe that's in C1.
206 
207~somlus yup
208 
209~ronreg ~somlus: physical commodity computers like x86 laptops don't
210        know whether a value in memory is a negative number, a positive
211        one, or something else, just from the bit pattern
212 
213~risruc then you might =match(C1,B1:b1000,0) to index the max position
214        in that range.
215 
216        (in C2)
217 
218~somlus yes
219 
220~ronreg other parts of the code will treat the same bit pattern
221        different if that part of the code knows its supposed to
222        represent a negative/positive/something else
223 
224        nock is no different
225 
226~risruc then you mgiht =index(b1:b1000,c2)
227 
228        of course, you don't need to do that in 3 cells. you can do it
229        in one, with: =index(b1:b1000,match(max(b1:b1000),B1:b1000,0))
230 
231        that maps almost perfectly to hoon.
232 
233        the difference is in hoon, usually you dont have parenthases.
234 
235~somlus Interesting ~roneg.. I will have to wrap my brain around the
236        code (which doesn't know the value/sign of the variable)
237        not needing to get this information from the computer that
238        theoretically does know the value / sign. (BTW risruc still
239        following your convo, just responding to roneg)
240 
241~risruc so you have to know what your runes expect and imagine them
242        yourself.
243 
244~ronreg ~somlus: someone used the term "casting" earlier. that's a term
245        of art in programming, that means, basically, to interpret the
246        same bit pattern a different way than was intended
247 
248~somlus Ah! ok got it. It's nice that excel very quickly tells you the
249        expected formats. But I do appreciate the analogy to multiple
250        lines vs one line, and the one line makes the "children" super
251        explicit and easy to see all at once.
252 
253~risruc i don't understand the whole language server system well, so
254        maybe someone else can fill in the blanks...
255 
256~somlus thanks ronreg. I'm going to copy this whole chat and literally
257        start to make a set of orienting instructions for the non-coder.
258        I'm actually already in touch with the Tlon guys who have a
259        considerable interest in just that.
260 
261~risruc but the hoon language server understands all of this, as you can
262        see by dojo barking at you when you try to type illegitimate
263        commands :)
264 
265~somlus ah true!
266 
267        that is actually comforting. I'd rather that then a 'syntax
268        error' of sorts at the end of all the effort.
269 
270~risruc ~somlus, as a non-coder myself, i'm happy to help out with that
271 
272~somlus cool risruc, BTW would have never guessed you are a non-coder.
273 
274~risruc i believe if you use vim, you can access the hoon language
275        server
276 
277~ronreg so, something that you can try typing in the dojo is: `@`-1
278 
279~somlus sorry what is vim?
280 
281~risruc well, i'm an engineer. and i've dabbled my whole life. but never
282        been a developer.
283 
284~somlus sure enough ronreg, it's a 1
285 
286~ronreg `` is hoon's syntax for doing a cast
287 
288~somlus ok copy that.
289 
290~risruc vim is the preferred cli text editor for most serious
291        developers. (i'm upsetting a minority of serious devlopers with
292        this comment)
293 
294~ronreg the @ represents a natural number, but -1 is a signed number
295 
296        so this means, take the bit pattern that represents a signed -1,
297        but pretend it's a natural number
298 
299        brb gonna lose my internet connection for a bit
300 
301~somlus got it. So I did this `@`-5 and got 9. I'll have to figure that
302        one out...
303 
304~ronreg try `@`-1, `@`-2, `@`-3
305 
306        and see if you can figure out the pattern
307 
308~somlus ok I will look into VIM
309 
310~ronreg also try `@`--1, --2, etc. -- is hoon's notation for + (it's -,
311        and then - again)
312 
313~somlus yes, got the patterns. I think I will have to hand do this in
314        binary to understand the conversions. How do you represent
315        negative numbers in binary?
316 
317~ronreg this is vim's website: https://www.vim.org/ . it's a text editor
318        that a lot of programmers like to use (I use it myself), but
319        your exact choice of text editor isn't super important
320 
321        programmers care about it the same way I imagine professional
322        carpenters care about specific brands of hammer or circular saw
323        or something :)
324 
325        ~somlus: there's actually multiple ways you can represent
326        negative numbers in binary
327 
328        but one of the interesting things about nock is that it doesn't
329        actually have a primitive notion of binary numbers at all
330 
331~somlus ok I've been using vscode and did install the hoon extension.
332 
333~ronreg it just has natural numbers and pairs
334 
335~somlus interesting...
336 
337~ronreg so, hoon chooses the convention that -1 is underlyingly 1, --1
338        is underlyingly 2, -2 is underlyingly 3, --2 is underlyingly 4,
339        and so on
340 
341        that obviously works, but nothing that says *only* that
342        particular convention of representing signed integers as natural
343        numbers works
344 
345~somlus so if it is not binary that leads to the logic behind the
346        conversion of -1 to 1 and -2 to 3 and --2 to 4, then what is
347        the logic behind that? just a "convention" rule and not a
348        calculation?
349 
350        unfortunately I am now at the point where my brain is whipped
351        around with the abstraction of negative numbers have alternate
352        representations. I'm trying to make it concrete so I can relate
353        to it, but perhaps that is not possible.
354 
355~ronreg it's a pretty straightforward way of solving the problem
356        of representing signed numbers when you only have unsigned
357        primitives
358 
359        another way you could do it is have a cell where [0 n] is
360        positive n, and [1 n] is negative n
361 
362~somlus that is much more concrete to me. (and here we mean that 1 being
363        false in hoon represents negative?)
364 
365~ronreg actually it's completely irrelevant that hoon treats 1 as false
366        for this
367 
368        it's just a convention
369 
370        I could also say [23 n] is positive n and [877 n ] is negative n
371 
372        I just pulled those numbers from thin air they don't mean
373        anything specific
374 
375~somlus oh ok
376 
377~ronreg by the way, it's worth noting that a lot of what we're talking
378        about is sort of general computer science concepts, and not
379        things that are specific to hoon or to nock
380 
381~somlus actually that is what I want - so that I can get oriented with
382        the specialist concepts and then apply them to hoon.
383 
384~risruc correct, as mentioned earlier by ~tirlyn:
385 
386        https://en.wikipedia.org/wiki/Two%27s_complement
387 
388~ronreg I think I mentioned that I'm just a beginner at hoon myself,
389        but I've been a programmer and student of computer science for
390        a long time, so some of these things that come up when learning
391        hoon are things I've seen before in other contexts
392 
393~risruc there are many ways to represent signed integers in binary
394 
395        that wiki article gives a decent rundown
396 
397~ronreg hoon uses a lot of idiosyncratic terminology for things that
398        have a "standard" technical terms in computer science
399 
400~somlus ok thanks for that article, will read it carefully.
401 
402~ronreg for instance, someone mentioned the concept of an "aura" earlier
403        - most people who study computer science will call that a
404        "type", and hoon is being sort of deliberately unusual in
405        picking another name for it
406 
407~rabsef it does sound cool tho, one must admit
408 
409~risruc that is a common enough critique, but i've come to think it's
410        not quite fair.
411 
412        urbit makes no bones about auras being types
413 
414~ronreg some people dislike the weird terms :)
415 
416~risruc however an aura is a subset of all types
417 
418        it is a very specific sort of type
419 
420~somlus yes the confusion for me is that since it is all new it's hard
421        to know where to look to find the definitive explanation in
422        plain english.
423 
424~risruc i.e., it is limited to atoms (numbers)
425 
426        there are more complex types which are not auras
427 
428~somlus ah yes. There are strings that are types as well, would that be
429        correct?
430 
431~ronreg ~risruc-habteb: that's absolutely a fair point, the people who
432        work on hoon sometimes pick these unusual terms deliberately
433        because they don't want to pretend that the hoon concept is
434        exactly the same thing as a more traditional CS concept
435 
436~risruc because they can't be represented by single numbers (they
437        require cells (binary trees of numbers)
438 
439~rabsef like a tape?
440 
441~ronreg ~somuls: there are two kinds of strings in hoon, a tape and a
442        cord
443 
444        actually, I might've misremembered the name of the 2nd one
445 
446~risruc i think that's correct. i've heard the same thing about saying
447        "cell" instead of "tuple"
448 
449~rabsef I think that's right ronreg
450 
451~risruc but a cell is a specific sort of tuple (it's a 2-tuple)
452 
453~ronreg ~risruc: I htink you could argue that a "cell" and a "tuple" are
454        TECHNICALLY different concepts, just similar ones
455 
456        anyway the nock documentation uses the word "cell"
457 
458~somlus getting back to types, I was guessing that a string was a
459        non-natural number type of type. what is an example of a
460        non-natural number type in other computer science realms?
461 
462~ronreg somlus: try typing `@`'a'
463 
464        in the dojo
465 
466~risruc an address book entry.
467 
468        or a recipe...
469 
470~ronreg a floating-point number
471 
472~risruc or, the format that chat-store in urbit uses to save all chats,
473        including metadata about time sent, sender, recipient channel
474 
475~somlus yes that's 97
476 
477~ronreg that is, a number like 2.3 that can have a decimal point
478 
479        somlus: are you familiar with ascii?
480 
481~somlus indeed.
482 
483~risruc no, actually floating points are still atoms.
484 
485        /auras
486 
487~somlus just typed in alt-97 and got a. so thanks that is clear
488 
489~risruc a list is a type which cannot be cast to an atom
490 
491~ronreg oh, non-natural-number type, I misunderstood
492 
493~somlus ah yes.
494 
495~ronreg anyway, at the dojo, something you put in 'single quotes' is a
496        cord, and this is also represented underlyingly as an atom
497 
498~somlus I'd love to go back to the `@`--2 portion of things. that
499        is still too abstract for me. I understand it now to be a
500        naming convention, a non-calculated way to represent negative
501        numbers. But I do not understand why they are interpreted as non
502        absolute-value positive numbers.
503 
504~risruc oh yeah, good example ~ronreg.
505 
506~somlus I see the pattern but do not understand from what it derives.
507 
508~ronreg specifically the atom you get by doing this is the ascii value
509        of each character concatenated together
510 
511~risruc getting back to that in one second ~somlus.
512 
513        and then a couple more threads i want to tie up before i check
514        out for the night.
515 
516~ronreg which is why `@`'a' (casting the cord 'a' to an atom) is 97
517 
518~somlus copy that
519 
520~ronreg 'aa' is 97*256^`+ 97*256^0
521 
522        which works out to be24.929
523 
524~risruc here, let me make this simple ~somlus (too late? ^^)
525 
526        the basic data type in hoon is the noun, *
527 
528~somlus perhaps! But if I retain 50% of all of this then i'm that much
529        closer..
530 
531~ronreg anyway, this is just like the `@`--2 thing - it's a convention
532        the people who implemented hoon chose because it made sense
533 
534~risruc it can be either an atom (number) or a cell (pair of nouns)
535 
536~ronreg the fundamental problem they were trying to solve was, how do
537        you represent a string of characters, or a signed integer, if
538        all you have are unsigned integers and cells
539 
540        and ultimatley anything else. everything you can represent in
541        hoon is at the bottom some combination of atoms and cells
542 
543~somlus ok I understand that there had to be some exotic solutions
544        / representation conventions to being constrained in the
545        fundamental representation of things.
546 
547~risruc OK, let me give you a couple more things that will help
548 
549~somlus I know you guys probably have to go and I should too. FYI I
550        did not understand how 'aa' is 97*256^`+ 97*256^0 works out to
551        be 24.929. My brain tries to work out that as a calculator and
552        doesn't quite knwo what goes where.
553 
554        ok cool risruc
555 
556~risruc `*`noun will cast any subject to an untyped noun
557 
558        `@`atom will cast any atom to an untyped atom
559 
560        `@ud`atom will cast any atom to an unsigned decimal
561 
562        play around with those and see their effects.
563 
564~somlus ok will do.
565 
566~risruc in many cases, all 3 will give you the same results. however,
567        you CANNOT cast a cell to @.
568 
569        there are two types of strings in urbit.
570 
571        'tapes' are single numbers which represent all characters
572 
573~master i'm late to the party here, but i'd like to correct ~tirlyn,
574        hoon's signed integers do not use two's complement
575 
576~risruc SHIT. recall.
577 
578        'cords'
579 
580        "tapes" are lists of each individual character's ascii code.
581 
582~master the nice properties of two's complement only hold for sized
583        integers, but ours are arbitrary-precision
584 
585~somlus ok will disregard twos compliment
586 
587        (still listening risruc)
588 
589~risruc ahh, ~master. that makes a lot more sense. i was trying to make
590        sense of the two's complement page wrt urbit's implementation.
591 
592        OK, so you can cast any atom aura to '@', including 'cords'
593 
594        but try to cast a "tape"
595 
596        `@`"tape"
597 
598        that's because tapes are not atoms. they are lists of atoms.
599 
600        we need to cast it up to raw noun.
601 
602        `*`"tape"
603 
604        `*`"tape"
605 
606        now, you can see at least, the reason that an aura is not a
607        general "type", but a specific subset of "types"
608 
609~somlus ok i'm going to copy this chat and study it a bit more and
610        try to come up with some good questions. I do appreciate this
611        conversation very much and many things are further along than
612        they were. There are some odd cloudy mental blocks still
613        - my mind isn't making the leap from concept to intuitive
614        understanding, but I do feel that much of it will click
615        eventually. Once it does, I will be able to put into a language
616        that a non-coder can understand a bit more efficiently than I
617        have been able to so far.
618 
619~ronreg there's a sense in which -1 in two's compliment is sort of
620        "infinity", -2 is "infinity -1 ", -3 is "infinity -2", etc.
621 
622        ~somlus: yeah, this stuff is not super-easy to wrap your head
623        around
624 
625~risruc by the way, back to our metaphor with excel an nested formulas..
626 
627~ronreg it takes time and practice
628 
629        lots of things in computer science and programming make more
630        sense after you've written code and sort of have a feel for what
631        problems you need to solve to do it
632 
633~risruc i had made this image to explain to another student, how the
634        code in walkthrough 1.3.1 is nested...
635 
636        however, keep in mind that after i made this, it was realized
637        that the code was not in good form.
638 
639        and it's been reformatted in the walkthrough.
640 
641~somlus I actually copied that to my computer risruc, sensing that it
642        could help me understand things eventually!
643 
644~risruc i should update that image with the properly formatted code.
645        forgot about that until now.
646 
647~somlus when you originally posted it
648 
649~risruc haha
650 
651        well the funny thing is, for someone who's handy with excel.. i
652        think the excel metaphor is better.
653 
654~somlus ok i'll check back here tomorrow. time for this tired brain to
655        turn in and ruminate a bit on this...
656 
657~risruc didn't realize that until using it just tonight. lol.
658 
659~ronreg cool, good night
660 
661~somlus yes, what I really need is a point of reference I understand to
662        get to something that is new. That helps. I'm thinking of this
663        from a "casio calculator" point of view and it is not adding
664        up at all. I'll get there. Thanks again for your time risruc,
665        roneg, master!
666 
667~risruc pax vobiscum
668 
669~ronreg no problem
670 
671        ~risruc: *tecum :)
672 
673~risruc that would mean "peace is with you", which I do home. but I
674        meant "peace be with you" ^^
675 
676        do hope*
677 
678        admittedly, my latin is poor. my apologies if i got that wrong.
679 
680        or, if you meant to respond rather than correct.. then i should
681        have just closed this window after my 'exit'. good night
682        friends.
683 
684~ronreg good night!
685 
686        ("vobiscum" is "with you (plural)", "tecum" is "with you
687        (singular)". catholic priests say "pax vobiscum" because they
688        are addressing the whole congregation, i.e. more than one
689        person)
690 
691~risruc naturally I was addressing the lot of you!
692 
693        but now that you spell it out, I see. gracias.
694 
695~tirlyn ~master thank you for noting that, please enlighten me, what do
696        we use as an encoding for signed integers?
697 
698~master hi ~tirlyn
699 
700        atoms (ie, in nock) are arbitrary precision
701 
702        unsigned atoms in hoon are just atoms @u, @ub, @ud, @uwhatever
703 
704        signed atoms in hoon are also arbitrary precision, @s and
705        friends
706 
707        two's complement is a scheme for fixed-with integers
708 
709        instead, we encode the sign in the low bit
710 
711        0 is positive, 1 is negative
712 
713        so the positive-sign representation of any unsigned atom is just
714        (mul 2 x)
715 
716        ie, (lsh 0 1 x)
717 
718        and the negative sign representation is one less than that
719 
720        i believe there's a name for this scheme, but i'm having trouble
721        finding it at the moment
722 
723        hoon has syntactic support for signed atoms
724 
725        --2 is positive, -2 is negative
726 
727        the signed functions are in the +si core
728 
729        2
730 
731        (abs:si -2)
732 
733        here's a gate (function) to test postive signed atoms:
734 
735        |=(a=@sd =(a (mul 2 (abs:si a))))
736 
737        and this will test negative signed atoms:
738 
739        |=(a=@sd =(a (dec (mul 2 (abs:si a)))))
740 
741        demonstrating the representation
742 
743        our code examples in chat aren't working quite right, but you
744        can copy those and run them in :dojo
745 
746        %.y
747 
748        (|=(a=@sd =(a (dec (mul 2 (abs:si a))))) -81)
749 
750        ^ like so
751 
752        ah, i remember, this is the "zig-zag" encoding
753 
754        notably used in protocol buffers
755 
756        https://developers.google.com/protocol-buffers/docs/encoding#typ
757        es
758 
759~tirlyn interesting! thank you.
760 
761~timluc going insane with hoon build stuff atm
762 
763        made a dev ship, then am doing a simple sanity check that keeps
764        failing
765 
766        But let me switch that to a more productive question: how can I
767        interpret syntax errors in Hoon better?
768 
769        For example, if I recursively call a function with something
770        like `$(j (dec j), a t.a)` and I forget the comma, I
771        get the error `ford: %hood: syntax error at [4 17] in
772        /~zod/home/0/gen/lists/hoon`
773 
774        does that mean that my error is in the 17th element of line 4?
775 
776~ronreg ~timluc: yeah that's what that [4 17] means
777 
778        that's documented somewhere
779 
780~timluc Cool thanks
781 
782~risruc note that that's not always where your error is!
783 
784        it's often the case that your error is before the indicated
785        point, but the compiler didn't catch it until then.
786 
787        for an easy example, a ?& conditional can take any number of
788        arguments. therefore it needs to be closed with ==.
789 
790        if you forget the ==, you'll see the end of the last line in
791        your error, because that's when the compiler gave up looking for
792        ==
793 
794~timluc Yeah, I've been playing around some today and got a more
795        intuitive feel
796 
797        for the error reporting--it's kind of growing on me
798 
799~rabsef somlus - 2 invites accepted - 2 new urbiters - smart developer
800        types. Hope they make a splash
801 
802        And your star bears fruit!
803 
804~somlus wow if you get lean on invites and are bringing in those sorts,
805        just let me know. Plenty more where that came from for quality
806        invites like that.
807 
808~rabsef Thanks - I'm trying to be discerning w/ the invites and get
809        people who will work on it or use it often. I have one other one
810        outstanding to a music producer for LoveCrypt records but he
811        hasn't started it up yet.
812 
813        I sincerely appreciate you letting me invite people
814 
815~somlus absolutely, please keep buildin' it up with them good folks.
816 
817        No time right now to pursue actual coding training (schedule
818        will open up a few hours from now) Now taking suggestions for
819        ideas for very (*very*) simple generators we can make here that
820        show basic (*basic*) principles. Before you make suggestions,
821        please once again read the name of the chat channel. We are the
822        hoontards. I was thinking that one could be a simple generator
823        that takes two inputs (real numbers), makes them into negative
824        numbers regardless of original sign, adds them together and
825        outputs the correct answer (also would be a negative number of
826        course). Another might be to take a real number as an input
827        and then output it with a variety of auras. Any other ideas
828        for exraordinarily easy initial code projects? Imagine you are
829        trying to teach code to your smart 8-year old child who is just
830        getting started. What projects would you give him/her to teach
831        them the very basic principles of something?
832 
833~rabsef Somlus - I have some hoon code I wrote previously walking
834        through the lessons online. These include flop (switching the
835        ordination of a list) and cypher (replacement cypher a=b, b=c,
836        and so on) and uncypher and I remember enough to be able to help
837        go back thru those.
838 
839        let me know if that's interesting
840 
841        My prior coding experience is Visual Basic and Excel formulas so
842        I'm definitely not "out there" in my comprehension
843 
844~somlus Smart move rabsef - those are very good basic examples.
845 
846~rabsef https://pastebin.com/0mXJntLe
847 
848        That's flop - pretty easy to walk through
849 
850        I added an additional "feature" that appends my ship name to it
851        lol
852 
853        Also I like flop because it shows 3 uses of tis. bartis, input.
854        tis/ (forget name) standard declared constant. tisbar empty
855        output declared variable
856 
857        my understanding of their function there may be a little limited
858        but it's functionally true
859 
860        I would be down w/ like a discord session per week where we go
861        through those lessons w/ someone smarter than me.
862 
863        I can take us through the first few but I get lost after the
864        cypher activity
865 
866~somlus ok cool. that sounds great. I will add that to the list.
867 
868~minder I have zero understanding of what in the world Hoon's purpose
869        is and/or how it connects to the experience of urbit. It's like
870        looking at math problems without understanding when/how I'd be
871        using the principles in the real world. Without understand that
872        relationship I'm totally lost.
873 
874~somlus yes it's very hard. these hoonsaints go off and so something
875        magical with their runes and whatnot, and then change happens.
876        Meanwhile, I'm struggling with very basic syntax. Risruc and
877        rabsef were super helpful to help me understand that, at the
878        outset, you need to understand the basics of syntax. In my
879        opinion that is nearly completely left out of the documentation
880        (for the noob). I am going to go to the Urbit IRL event in SF
881        tonight and will talk to the head of documentation about how we
882        wrap our heads around teaching to the very new individual.
883 
884        The good news is that, from what I understand, if you can make
885        your way around excel, you have the 'what goes where and what
886        it does' basic skills to code. Confirmed this with a masters in
887        Computer Science IRL yesterday too. So I think we should ask a
888        hoonsaint to take us ridiculously meticulously through basic
889        code until we understand what goes where and why. Then we can
890        start to expand into complex strcutures.
891 
892        *confirmed this with someone who has masters in CD...
893 
894        *CS (dammit)
895 
896~minder "How does this Nock thing go from a tshirt to making urbit my
897        own Google Drive available on phone & laptop?" Simple question.
898        X^D
899 
900~somlus just a simple question. Can someone please tweet the simple full
901        answer?
902 
903~risruc sweet somlus. I'm still interested in helping with that project.
904        glad the excel metaphor helps.
905 
906        let me know how those conversations go tonight
907 
908~minder X^D
909 
910~somlus will do! ~lodleb-ritrul is the head of docs and he will be
911        there. (only a bit more time left here until I gotta bounce,
912        but..) I am trying to wrap my brain around 1) identifying
913        basic CS terminology 2) drawing some parallels between such
914        terminology and Urbit-specific terms, and most importantly
915        3) figuring out how much to 'dumb down' these terms so that
916        they can be consumed by the common person without creating
917        misunderstandings later. The thing about CS terms is that they
918        need to be super precise in order to be accurate regarding the
919        tiny niche of the world they are trying to describe, otherwise
920        you make mistakes later. But if you have to convey the exact
921        precise nature of the exact definition & its nuances, you lose
922        the student. So somewhere is the sweet spot for dumbing down
923        the concepts to their core, but also providing the complete
924        definition so that people can dig deeper when they are ready...
925        at least that is what I have for now.
926 
927~risruc I totally agree with your characterization of the problem and
928        the challenge. currently I am leaning towards less emphasis on
929        mapping to other CS terms, in favor of more of a blank slate
930        approach.
931 
932        that will be most helpful for the total newcomer anyway. and
933        with proper explanation, should suit the convert developer as
934        well.
935 
936        of course when you're speaking to a convert, it is helpful to
937        be able to explain at least the closest approximate mapping of
938        concepts and terms, with the required caveats.
939 
940~minder Communicating complex ideas to non-understanders is my jam. I
941        love it. In terms of computer science, I had enough knowledge
942        with html to script-kitty my first site together. Been building
943        wordpress sites for a long time using page builders. Can do
944        DNS editing etc. So I'm competent to get things done if I have
945        a walkthrough to follow. But nothing in terms of creative
946        expression within any kind of computer input dynamic.
947 
948~somlus ok cool we're pretty similar. maybe this weekend we can find a
949        few minutes to go through a simple "take two inputted numbers
950        and add them together and produce output" code so we can start
951        to get comfortable with the format and simple competency. We
952        might be creating Hoon 001!
953 
954~risruc one really gets the feeling of putting the band back together
955        here.
956 
957~minder Hoon 001; I love that.
958 
959        I'm very much a 'why-first' kind of thinker vs a 'what-first.'
960        "What you do here is a + b = sammoflange hoopty bop" is what I
961        hear. But if you go "Why is it this way? Because we need X to Y
962        as a function of B" I'm on board.
963 
964~risruc Hoon 001 -- The Blind Leading the Blind
965 
966        alternate subtitle -- Getting the Band Back Together
967 
968~lapfur Here is a cool talk by Richard Feldman, one of the
969        more prolific and engaging Elm programming educators:
970        https://youtu.be/G-GhUxeYc1U (Elm is like a functional
971        programmers' version of JavaScript) He basically summarizes
972        educational theory in easy steps for those that teach
973        programming.
974 
975~risruc okay. I really need to not text and drive. lol. I'll talk to you
976        later.
977 
978~lapfur Looking forward to following this chat while I learn^^
979 
980~somlus cool thanks lapfur!
981 
982~minder I read that as Richard Feynman and I was all "holy cow he did
983        programming TOO?!"
984 
985~lapfur Lol
986 
987~somlus Ok minder we'll mix what and why - that's good. Keep in mind
988        that there is a tendency with super complex stuff to let
989        discussion trail off in all sorts of interesting directions.
990        So as long as we can keep the what/why/how all focused I think
991        we'll be great here.
992 
993        (interesting but educationally inefficient directions)
994 
995~minder I'm 4 minutes into this video and I'm vibing so hard with what
996        he's saying. X^D
997 
998~somlus cool will watch it soon
999 
1000~minder His "HTML vs HTML msg" difference question/explanation
1001        experience is exactly what I feel when people talk about Hoon.
1002 
1003        Frame the Goal: That's the why! Why do we do this. ~lapfur,
1004        thank you so much for posting this video.
1005 
1006~somlus yeah that video is pretty much 1000% correct so far
1007 
1008~risruc yeah I just started video. great so far. thank you.
1009 
1010        n.b., not driving any more. lol.
1011 
1012~minder ~risruc; I can't get landscape to load on my phone, so I can't
1013        drive&urbit. . . yet.
1014 
1015~lapfur As I'm going through the docs, it would be cool to have a sort
1016        of mnemonic translator to help us learn the different kernels
1017 
1018        For example, (these are guessses that I"m ma
1019 
1020        I think the original designers made when naming things:
1021 
1022        Behn = Big Ben, like the famous clock that is the timer vane
1023 
1024        Nock = nocking on the binary tree
1025 
1026        Ford = the assembly line build system like the Ford cars
1027 
1028        etc.
1029 
1030        Well, better written than that, but hopefuly you get the idea
1031 
1032~minder https://www.amazon.com/dp/0982370849
1033 
1034        A short book I wrote about mnemonic techniques like linking &
1035        peg systems. X^D
1036 
1037        Having associations to stuff you already know is a great way of
1038        making new information 'sticky'
1039 
1040~lapfur You weren't kidding when you said 'this is my jam'^^
1041 
1042~minder =^D
1043 
1044~ronreg ~minder-folden: in the 80s richard feynman was involved with his
1045        son's computer startup
1046 
1047        as a technical consultant
1048 
1049        https://www.youtube.com/watch?v=EKWGGDXe5MA[embed]
1050 
1051        ~minder-folden: in the 80s richard feynman was involved with his
1052        son's computer startup
1053 
1054        feynman was a smart guy
1055 
1056~somlus This question is pretty simple so I will put it here. Can
1057        someone explain the basics to me about |mount and |commit %home
1058        ? 1) What is it about linux structure or working within linux
1059        (Ubuntu) that makes it necessary? 2) when you do the |mount
1060        command, is it basically copying the files from the working
1061        / operating format to a regular file directory format, but
1062        then also retaining the original data unchanged in the working
1063        format? and 3? when you |commit %home are you basically doing
1064        the opposite, with a program that has been watching for changes
1065        in the mounted directory and then re-injects the changed files
1066        into the live operational format? Please set me straight on any
1067        misunderstandings of this supposition. thanks!
1068 
1069~master ~somlus, short answers:
1070 
1071        |mount sets up an external mirror of some %clay desk
1072 
1073        |commit injects the files in that external mirror back into
1074        %clay, resulting in a new revision of the relevant desk if any
1075        of those files have changed
1076 
1077        so, yes
1078 
1079~somlus Cool, makes sense to me now. Is this a common thing in Linux or
1080        fairly specific to urbit?
1081 
1082~master the internal structure (%clay's state) and external structure
1083        (linux filesystem) bear no resemblance to each other, so
1084        transformations are needed in both directions
1085 
1086        our implementation is very basic, the simplest thing that could
1087        work
1088 
1089        it used to be automatic, but there were subtle bugs
1090 
1091        filesystems are hard ...
1092 
1093~somlus and critical. Ok that makes complete sense to me thank you.
1094 
1095~master the details are specific to urbit, but exposing some data
1096        through a filesystem interface is very unix-y
1097 
1098        linux exposes it's process true through something called procfs
1099 
1100        and FUSE is a general-purpose way to expose arbitrary data as if
1101        it were a filesystem
1102 
1103        urbit used to have a FUSE implementation for browsing the
1104        namespace
1105 
1106        just a demo, but it'd be nice to bring it back and make it real
1107 
1108        that should be "process tree"
1109 
1110~somlus which should be "process tree" (much of this is new to me). Also
1111        I'm going to ask a f/u question on u-h because it might benefit
1112        from multiple eyes on it /responses. Thank you for this info!
1113 
1114~risruc s/"process true"/"process tree"
1115 
1116~somlus ~rabsef-bicrym if you have a link to the cypher program you
1117        wrote can you post it here?
1118 
1119        This outstanding notes index for Hoon school was recently
1120        posted. I can't remember by whom! Anyone who can, please post
1121        that ship name here
1122 
1123        https://www.notion.so/Hoon-School-Notes-Index-e84498f0f38f4948b1
1124        6d5bab4fd5dece
1125 
1126        Also helpful for people super comfortable with Windows but new
1127        to Linux, article which is a friendly read on how unix/linux
1128        works: https://neilkakkar.com/unix.html
1129 
1130~rabsef https://pastebin.com/gGqMK0Wm
1131 
1132        If I'm reading that right, it's after the second step where
1133        you're supposed to shift twice.
1134 
1135        I need to pick that back up. I'm going to start tomorrow - thank
1136        you for inspiring
1137 
1138        I'm going to finish the Hoon School 101 that I'm in - I stopped
1139        because I couldn't get to the dang notebook but that's been
1140        resolved for weeks
1141 
1142        I can get it done and then I'll share my full note set from that
1143        and start going back through the tutorials
1144 
1145        I also want to map all the potential error flags you can see in
1146        urbit - I'll try to fit that in
1147 
1148~somlus wow!
1149 
1150        that's a lot
1151 
1152~rabsef I'm going through the hoon school notebooks and taking notes
1153        on each reading. Almost done w/ week one (which I already
1154        completed, but then got locked out so I gave up)
1155 
1156        If it would be nice to share I would be happy to - a lot of it
1157        is covered on teh website but maybe my repetition of it would be
1158        helpful
1159 
1160~master ~somlus, those are ~fabled-faster's notes
1161 
1162~fabled feel free to copy them if you use notion - i think you can
1163        actually rip the entire directory of notes
1164 
1165        if you don't, I'm 90% sure the text you can copy turns into
1166        markdown
1167 
1168~somlus I thought it might be fabled. thank you - those are brilliant!
1169        It give me new verve to go make another attempt.
1170 
1171~rabsef Can I try sayin hi to one of u using : ?
1172 
1173~somlus Yes you bet - should I just watch the dojo?
1174 
1175        FYI I'm going to try to change a bios setting on my computer, so
1176        I'm doing a restart.
1177 
1178~risruc i'm looking at cli now, ~rabsef
1179 
1180        if you want to hi me.
1181 
1182        ; ~rabsef-bicrym is your neighbor
1183 
1184        hi ~rabsef-bicrym successful
1185 
1186        |hi ~rabsef-bicrym
1187 
1188~rabsef sorry - I stepped away for a second
1189 
1190        lol
1191 
1192~somlus I'm back.
1193 
1194        that's what happens if you try to |hi in chat
1195 
1196~rabsef trying now - yes somlus
1197 
1198        risruc - it should just have sent
1199 
1200        sweet
1201 
1202        I'm doin something other than |hi tho - I'm using : and in the
1203        form of :@p/app &helm-hi @t
1204 
1205        so like :~rabsef-bicrym/hood &helm-hi 'hi'
1206 
1207        dot poast
1208 
1209~somlus I see nothing special in chat-cli.
1210 
1211        -------[15.635] showed up after you said you sent it
1212 
1213        other than that looks like normal stuff
1214 
1215~risruc welcome back ~somlus.
1216 
1217        with your wax, i wane.
1218 
1219        time to get ready for bed.
1220 
1221~rabsef didja get that som?
1222 
1223~somlus Nope. What's it supposed to look like?
1224 
1225~rabsef I crashed
1226 
1227        It should look like a hi
1228 
1229        I seg faulted tho - let me try again
1230 
1231        I just tried - it read 'test test one two three'
1232 
1233~somlus Nothing. Have the window open in chat-cli, not message of that
1234        sort received.
1235 
1236~rabsef isruc
1237 
1238        That was supposed to read "I was able to do it with our friend
1239        risruc
1240 
1241        it may just be latency or i may be doing it wrong
1242 
1243        I'll try a normal hi
1244 
1245~somlus I saw this is a normal hi
1246 
1247        oh i see test test one two three up there after all
1248 
1249        it has your ship name then colon then the text
1250 
1251~rabsef NICE!
1252 
1253~somlus saw this is the other way
1254 
1255~rabsef that's using format :~sampel-palnet/hood &helm-hi 'text'
1256 
1257        in that I'm sending a &helm-hi message with that tape to your
1258        hood rather than relying on Hi to do it
1259 
1260~somlus Great, got both
1261 
1262~rabsef I don't really understand &helm-hi but it's cool
1263 
1264~somlus very cool. Ok won't be super close to the screen for the rest of
1265        tonight. Great that it worked. Gnite all.
1266 
1267~ronreg ~somlus-savlev: thanks for posting the hoon school notebook
1268 
1269~somlus Wish we could pin posts - I would definitely pin one with that
1270        in it.
1271 
1272~fabled yeah, pinning would be nice
1273 
1274~rabsef ~somlus if toure doung a call this weekend jd love tk join.
1275        aorey typos on phone and just wome up. fingers no work
1276 
1277~somlus will have you in for sure rabsef...
1278 
1279~rabsef I'm excited. Need voip for urbit!
1280 
1281        and sorry for double post in 2 contexts - I have resolved to try
1282        better w/ separation of chats in the new year
1283 
1284~somlus No worries - I did the same! I'd say post away we all know this
1285        is a challenging interface in terms of remaining current, etc.
1286 
1287~fabled ~rabsef - i too, am excited for most communication types to be
1288        afforded for in urbit
1289 
1290        it’ll be further out but we’ll get there eventually
1291 
1292~risruc one thing i don't understand, perhaps because i dont understand
1293        enough how computers work...
1294 
1295        is how will urbit handle real-time/continuous tasks
1296 
1297        like voip, video chat, live gaming
1298 
1299        since it is necessarily single-threaded computing.
1300 
1301        e.g., if you ask your urbit to calculate something (say
1302        1,000,000 factorial) and it hangs for some number of seconds
1303 
1304        does that mean you can't be sending or receiving any packets in
1305        the mean time?
1306 
1307        again, perhaps the question is silly..
1308 
1309~rabsef I /think/ that though it's single stream, there is currently
1310        some need to 'try again' for a period of time, but I see your
1311        point. Additionally, it may be that urbit is just the identity
1312        layer to a product like voip that sits at a higher layer?
1313 
1314        I don't miss traffic when I crash, to the best of my knowledge.
1315        It is delivered next time my urbit is available
1316 
1317~fabled we need someone like ~palfun to explain how this is possible
1318 
1319        all I know is that VoIP and streaming to a degree will be
1320        possible, but I forgot by what mechanisms it'll be possible
1321 
1322        you're correct in thinking that at the present time, tasks of
1323        that sort are not possible
1324 
1325        i forget if we have any urbit engineers/infrastructure team in
1326        here
1327 
1328        nvm, there are a few in the backscroll
1329 
1330~risruc will repost in /u-h, now that california is waking
1331 
1332~somlus Yes it does seem that something will have to sit on top of
1333        urbit ~risruc. I had never though of this until now but i seems
1334        self-evident. It will request data from the ship but not have
1335        access to it without permission.
1336 
1337        because for live chat it doesn't need to go through the urbit.
1338        The ship can set up a P2P and only be used for the key that sets
1339        up the connection, I think?
1340 
1341~nidsyp That's a good question ~risruc. How is multiprocessing
1342        implemented. I like to think of moons being used to that. But
1343        I'm not sure that's practical because they are rather heavy
1344        height.
1345 
1346        But maybe arvo will have something similar but much lighter
1347        weight.
1348 
1349~hastuc https://github.com/urbit/proposals/blob/master/005-asynchronous-
1350        multiprocessing.md
1351 
1352~risruc thanks hastuc. that's an interesting read.
1353 
1354~nidsyp thanks ~hastuc.
1355 
1356~somlus Wow hastuc, that is a real commitment to an idea. Well done.
1357        Immediately when you stated the problem it resonated with me. I
1358        am assuming they can create moons with different requirements
1359        from the traditional 2GB of memory
1360 
1361        Or they could create a category of ship called a satellite that
1362        is specifically suited to computational tasks, that the planet
1363        can auto-spawn and auto-destroy, and can only communicate with
1364        the planet or perhaps with one outgoing authorized entity,
1365        and the results of that "life" are then incorporated to the
1366        deterministic data set, somewhat akin to closing out a block on
1367        a blockchain (but only loosely so - just describing the keeping
1368        of data in a larger record here).
1369 
1370        The only benefit of creating a new type is that everyone knows
1371        it is only for competition, is expected to be destroyed, and has
1372        zero authority.
1373 
1374        computation (not competition). needs coffee.
1375 
1376        Also, I should read the thread next time since I see that about
1377        5 1/2 hours ago you mentioned something very similar.
1378 
1379~risruc haha, not a problem. good minds, et c.
1380 
1381~hastuc hey don’t thank me, thank ~rovnys-ricfer next time you see him,
1382        he wrote it
1383 
1384        ~solmus I didn’t write it haha ~rovnys-ricfer did
1385 
1386~somlus ah!
1387 
1388        misunderstood.
1389 
1390~nidsyp thanks for sharing it though
1391 
1392~ronreg I like hte idea of calling a computation-specific moon a
1393        sattellite
1394 
1395~somlus I regretted saying it had zero authority. It has the specific
1396        level of authority and data access that the planet gives it. It
1397        can only see a very narrow slice of the ship's file sert.
1398 
1399        set.
1400 
1401~rabsef thst is your joint learnign sesh
1402 
1403        somlus did i miss it?
1404 
1405~somlus no sir! I sent a message for you in minder's party and told
1406        you that today got away from me. When works for you tomorrow
1407        (sunday)?
1408 
1409        I'm guessing you will be going to church some of the day, but
1410        I'm 100% flexible so I'll plan it around your schedule.
1411 
1412        Just let me know good times and time zone and I'll work it
1413        around your schedule.
1414 
1415~risruc if I'm able to be around, I'd love to. but family party tomorrow
1416        evening. not free until later night. more free after Christmas
1417        for the next Meetup.
1418 
1419~somlus Ok we'll figure it out!
1420 
1421~hanwed Testing a moon (I will likely use moons when we are learning
1422        code - you'll know it's me because it ends in ~somlus-savlev
1423 
1424        Obviously if I trash a moon I"ll use another
1425 
1426~somlus .
1427 
1428~risruc wow, ~somlus, brutal.
1429 
1430        I know you have 4 billion of them
1431 
1432        but think of the children.
1433 
1434~somlus haha! oh, man. I didn't think of the kids! what have I done??
1435 
1436~rocted moon testing (will code from this moon later)
1437 
1438~waldev or this one
1439 
1440~risruc i absolutely love the devmoons
1441 
1442        i'll ahve to set some up.
1443 
1444        much nicer than comets or fakezods
1445 
1446~somlus yeah, It's nice to know you have a fully working urbit that you
1447        are testing. No need for waiting for comet mining, and a fakezod
1448        is of course not updated.
1449 
1450        What I do is just |moon over and over again like 10-20 times,
1451        then copy that whole section, remove the \/ and the xtra spaces
1452        and you voila, you have like 10-20 urbits ready to spawn. I use
1453        the -G setting and copy/paste the keys in there and they spin
1454        up. Easy, peasy.
1455 
1456~rabsef sorry I didn’t see your other message. we need DMs. i have
1457        tickets to a show at (all times mountain) 2pm today, and
1458        assuming it lasts 2 hours and drive time probably 6PM I’d be
1459        available conservatively.
1460 
1461        test
1462 
1463        what are you thinking of covering today? I was thinking about
1464        it and I think maybe just reviewing the noun chapter from the
1465        guides and do the first banal assignment from Hoon school. it
1466        covers the absolute basics of input and output in Hoon programs
1467 
1468        those took forever to post. sorry
1469 
1470~somlus Totally cool le
1471 
1472        let's plan on 5pm PST 6PM MST to get started. I think you will
1473        be shocked at how early of an understander I am for hoon. I
1474        need to begin at the beginning beginning, meaning that I do
1475        not fully even understand the syntax. Fortunately I know excel
1476        to an intermediate level, so I can easily have excel metaphors
1477        as examples. Actually i was thinking of creating super simple
1478        spreadsheet "programs" in excel and then trying to do the
1479        same thing in hoon. TBH I can't even understand the syntax of
1480        variables, user input and (now using hoonspeak) children. So a
1481        simple "let the user input two numbers and then tell them the
1482        sum of those numbers" would be where I would begin. And then
1483        from there I think we go through the hoon components of the
1484        exercises, like the create list exercise and so on.
1485 
1486~rabsef I'm going to try to write that quickly and see if I can be
1487        prepared and then I'll know what elements we need to refer to
1488 
1489~somlus Cool. Also I cannot recall who posted this video (maybe
1490        it was you rabsef?), but the first 8 minutes accurately
1491        spell out the dilemma and the beginning of the approach for
1492        simple, directed learning at the orientation level of things:
1493        https://www.youtube.com/watch?v=G-GhUxeYc1U&t=729s
1494 
1495~rabsef in fact that example is in the syntax of the correct rune
1496 
1497        so it's pretty easy lol
1498 
1499        shit yah I did it
1500 
1501        lol
1502 
1503        Hey - I did it - and my |commit took forever
1504 
1505~somlus yeah!!!
1506 
1507        interesting that your commit took forever. had a long time
1508        passed between your |mount and |commit?
1509 
1510~rabsef test
1511 
1512        yayaya im workig and i got the function programmed for later.
1513        recommend all do a |commit %home before we go just in case you
1514        have as much uncommitted as i did
1515 
1516        took a long time
1517 
1518        somlus, about to start my thing but - are you familiar eith
1519        editing files in linux? because to make a generator thst is
1520        effectively a named function in hoon, its much esdier to do it
1521        in a text editor then commit thsn to do it in dojo directly
1522 
1523        forgive typos - big thumbs and small phone
1524 
1525        in dact i basically dont do work in dojo cuz you need to be
1526        accurate which isnt my forte
1527 
1528~risruc also, ~somlus, i assume that you're not working from the machine
1529        that runs your urbit.
1530 
1531        if so, are you working from mac or pc?
1532 
1533        is that correct?
1534 
1535        you'll want to know how to post files with scp.
1536 
1537        if you're on windows, you can use pscp (comes with PuTTY)
1538 
1539~tocref Hello Friends –– I finished hoon 101 and am starting on hoon 201
1540        –– I have previous programming experience and I can def try to
1541        help out w any questions anyone here might have
1542 
1543~somlus ok got it.
1544 
1545        Can you simply re- mount? I'm sure I don't have too much in the
1546        mounted directory that is relevant and I'm ok with overwriting
1547        it.
1548 
1549~rabsef mountint puts files in unix, commiting reads them from unix to
1550        ship
1551 
1552        so remounting isnt what we want. intermission
1553 
1554        this might be slightly longer than i thought...
1555 
1556        unexpected - because its a children friendly program. steange
1557        choice to make it like 3 hours
1558 
1559~somlus how odd to make it 3 hours, for kids? That seems like a bit out
1560        of touch of them. Unless it's a holiday classic...
1561 
1562        ~rabsef I think the easiest solution until we understand this
1563        further is to launch moons and hoon off of them?
1564 
1565        (BTW I know you won't reply until some time later b/c you're at
1566        the event)
1567 
1568~tocref IIRC There's a way to launch a testnet comet that you can use
1569 
1570        Gonna try to find it in the docs
1571 
1572~somlus Question tocref - on https://packages.ubuntu.com/bionic/i3 there
1573        are lots of architectures to choose from. I recall that often
1574        things in ubuntu are downloading amd64 even though the hardward
1575        is actually an intel processor. Can you tell me whether I should
1576        download the amd64 or the i386 version to use?
1577 
1578~tocref https://urbit.org/using/develop/#creating-a-development-ship
1579 
1580        One of these isn't actually connected to the broader urbit
1581        network but you can still run hoon in it
1582 
1583        Which is convenient because it's still technically possible to
1584        break a ship trying to run a broken hoon program
1585 
1586~somlus ~tocref check out my post from almost exactly 5 hours ago.
1587        I like moons and spawn a ton of them and then use those for
1588        development. You can also comment as your moon and people will
1589        still know it's you since they are connected.
1590 
1591~tocref somlus –– I'm not 100% sure on how to run a linux VM on windows
1592        because I currently use a mac and previously just used linux for
1593        like 7 years lol
1594 
1595~rocted Still somlus here....
1596 
1597~somlus good to know, no worries tocref.
1598 
1599~tocref If you're trying to download a copy of ubuntu, unless your setup
1600        is really weird I'd download the latest LTS image from here
1601 
1602        https://ubuntu.com/download/desktop
1603 
1604~somlus No i'm trying to download i3 from within ubuntu. There are lots
1605        of choices and it must be the i386 one
1606 
1607~tocref Ah, is apt-get not working?
1608 
1609        Normally apt-get is supposed to figure out this stuff out for
1610        you
1611 
1612~somlus good point. I should just do it that way, thank you for the
1613        reminder.
1614 
1615~rabsef ok I’m out and heading home. cutting it marginally close but
1616        will be there by six Denver shit traffic bloody allowing
1617 
1618~somlus please take your time. it's pretty much you and me and i'm in
1619        for the night. will standby
1620 
1621~rabsef I'm here finally!
1622 
1623~somlus Kaboom! I think for now we should be fine with just us. If
1624        we get into questions that are hard to answer, then we will
1625        announce in urbit-help we are doing this and see if any good
1626        knowledgeable souls show up to help us out. Sound good?
1627 
1628~rabsef I made, in advance, a documented hoon to add two atoms and
1629        explain everything going on - it's very rudimentary but it's how
1630        I started to understand what creating hoon files does and how to
1631        do it
1632 
1633        Sounds good - how do you want to start?
1634 
1635        I think we should talk about the specific problem you're trying
1636        to solve (add two numbers) and what that means in urbit
1637 
1638~rocted ok cool. let's start with that! I'm going to talk from my moon
1639        from now on. will quickly mount it
1640 
1641~waldev oops wrong moon. that one doesn't have access to minder's party
1642        so I'll use this one.
1643 
1644~rabsef this covers 1) what is a number 2) how does urbit handle numbers
1645        3) how do you create hoon files and 4) what runes are relevant
1646 
1647~waldev Cool. ready!
1648 
1649~rabsef I'm going to pretend like I know what I'm talking about and that
1650        will help me solidify the knowledge ify the knowledge I already
1651        have
1652 
1653~waldev How would you like to share it?
1654 
1655~rabsef had a stroke there - you get the point
1656 
1657        https://urbit.org/docs/tutorials/hoon/nouns/
1658 
1659        we should start here. Everything, it turns out, is a number in
1660        urbit
1661 
1662        all text and so on are numbers. Any output that isn't a number
1663        (save tapes and cords but we'll get there) can be freely "cast"
1664        into a different aura
1665 
1666~waldev got and printed as well.
1667 
1668~rabsef but first lets talk about how the numbers are stored
1669 
1670~waldev fortunately that part I understand already.
1671 
1672        Great!
1673 
1674        I need that.
1675 
1676~rabsef ah ok well then lets move on. Do you feel comfortable w/ the
1677        binary tree notation and concept?
1678 
1679~waldev yes I do except for one question on that. I can't fully
1680        understand why such a complex system was required. I think that
1681        overall it will be easiest for me to use excel as an analogy. In
1682        excel, it's so easy to see where data is, and where data can be
1683        inputted and outputted because of the grid. Items are in cells,
1684        that are right there in front of our, A5, C2 and so on. Is this
1685        binary tree a means of naming different components to a one-line
1686        depiction of code, which must be depicted in the text of a
1687        terminal? Or why would they come up with something so complex?
1688 
1689~rabsef The important part to remember is that all cells are read from
1690        the "right"
1691 
1692~waldev yes fortunately I also understand that and also how the system
1693        will remove excess brackets.
1694 
1695~rabsef That's a good question - and yes. Because everything is read
1696        from the right when not explicitly bracketed, you're able to
1697        give a numerical value to the specific position in a tree
1698 
1699        e.g. +1:[22 [33 44]] as in the example, shows the whole thing
1700 
1701~waldev (incidentally one thing we might consider is documenting the
1702        'take away' messages like that one somehow so that future peeps
1703        can skip to the take home messages.
1704 
1705        yes indeed
1706 
1707        and +2 is 22, +3 is [33 44} correct?
1708 
1709~rabsef all the way down the number-names are consistent no matter if a
1710        specific branch of the tree exists
1711 
1712~waldev +6 would be 33
1713 
1714        I think.
1715 
1716~rabsef and +5 is nonexistent because 22 is a leaf
1717 
1718        precisely
1719 
1720~tocref Think the main reason it's so unique (most other programming
1721        languages don't store everything as a binary tree lol) is to
1722        allow it to handle state in a purely functional context (correct
1723        me if i'm wrong)
1724 
1725~waldev good new vocabulary. ok so leaf means a "dead end" in the
1726        bracketing system?
1727 
1728        welcome tocref!
1729 
1730~rabsef I'm sure you're right ~tocref - I have a limited understanding
1731        anyway so please don't take me as the expert
1732 
1733        A "leaf" is technically a noun that doesn't have any "children"
1734 
1735~tocref It's hard to explain what this means without writing a couple
1736        paragraphs on how other programming languages work and I'm not
1737        experienced enough to articulate it lol
1738 
1739~rabsef so yes, but also we should use that language I think
1740 
1741~somlus interesting tocref.
1742 
1743        ok cool. we'll use that language, but part of what I need as
1744        a beginner is boundaries on these definitions. I need to know
1745        both what they are and what they aren't. I'll probably keep
1746        describing in common terms until the hoon terms become more of a
1747        second nature.
1748 
1749~rabsef Yes - sorry - didn't mean it as a correction, just trying to use
1750        the language to again solidify it for me
1751 
1752~tocref Hoon's terms are like, also different from a lot of the usual
1753        terms, which is helpful sometimes and absolutely baffling in
1754        others
1755 
1756~somlus copy that, rabsef
1757 
1758~rabsef So, what we want to do is add two "nouns" which are "atoms"
1759        (because it's more complex to add cells). There's technically
1760        already a function that does this
1761 
1762        our atoms will be leafs, as (I think) all atoms are leafs
1763 
1764~somlus ok good rabsef thanks.
1765 
1766~rabsef i.e. trivial (non complex) holders of space in a binary tree
1767 
1768        w/ values
1769 
1770        if you try (add 2 3) in dojo, it should give you 5
1771 
1772        we want to recreate that behavior
1773 
1774        Are you familiar w/ the runes at all?
1775 
1776~waldev ok grate done, and thanks.
1777 
1778        yes I have them printed out and flashcards to learn the
1779        phonetics. But I do not at all understand how they work. I try
1780        to think of them like excel commands
1781 
1782~rabsef they are, sort of, incidentally I'm super into excel so this
1783        will work well
1784 
1785~waldev where you have a certain number of things you can separate
1786        by commas and put in the parentheses in excel, each of those
1787        separable things being "children" here
1788 
1789        ok great. i'm a lower intermediate level user of excel
1790 
1791~rabsef Let's just jump to what we need - I happen to know, but we need
1792        a core
1793 
1794        We want a core that allows us to do two things - take input and
1795        act on it
1796 
1797        thankfully there is that core, bartis
1798 
1799~tocref I'm about to disappear for a little bit; I'm going to start on a
1800        hoon 201 assignment (again, as my HD got wiped a while back)
1801 
1802~waldev so is a core like a command in excel?
1803 
1804~rabsef The syntax for bartis is (bartis) <specification of input> <hoon
1805        to act against it>
1806 
1807        lol toc you should be teaching this
1808 
1809~waldev ok got it
1810 
1811        uderstood toc thanks
1812 
1813~rabsef Explaining what a core is (battery paylod) is a little beyond my
1814        understanding. Basically I /think/ the battery is what to do and
1815        the payload is what to do it with
1816 
1817        so yes
1818 
1819        that definition comports w/ our "syntax" for bartis |= a=spec
1820        b=hoon
1821 
1822~rocted ok thanks. Got it.
1823 
1824        (switched moons sorry)
1825 
1826~rabsef considering that other hoon runes are part of hoon, I think a
1827        core could have a core inside of it so like excel again
1828 
1829        so, we know we need |= and we know we have (add a b) already as
1830        a function in hoon
1831 
1832~rocted ok i'm with you
1833 
1834~rabsef that gives us |= <a=spec> (add x y) where x and y are somehow
1835        going to come from a spec
1836 
1837        recalling that |= requires a=spec b=hoon, add is our b
1838 
1839        We already also know what specification we want to give to our
1840        inputs
1841 
1842        They aren't dates or hexidecimal or anything, they're just atoms
1843 
1844~rocted ok hold on just a sec there. Can you tell me what <a=spec> is
1845        and why it is there?
1846 
1847~rabsef for arguments sake, lets say we want to give the computer a cell
1848        to be added [asub1 asub2]
1849 
1850        yes sorry
1851 
1852        a=specification meaning a is input values specified in a type
1853 
1854~rocted ok thanks. so it tells you whether x and y are going to be in a
1855        binary, decimal or other notation, even though no matter what
1856        they are going to be real numbers?
1857 
1858~rabsef just like in excel, like suppose the right argument =RIGHT(cell
1859        (a),starting point)
1860 
1861~tocref Before I leave, when you get some sort of code written make sure
1862        to put into a pastebin or a github gist and drop the link in the
1863        chat, easiest way for me to critique IMO (Don't think urbit chat
1864        lets you post code snippets)
1865 
1866~rabsef right is b=hoon, a is the interior of the parens
1867 
1868        I will post the pastebin for sure
1869 
1870        let me ask palfun something quickly
1871 
1872~rocted ok
1873 
1874        going to study this a bit anyway
1875 
1876        thank you tocref will do
1877 
1878        rabsef, I know you have issues with your ship, so no rush, but
1879        when you come back:
1880 
1881~rabsef ok back to this
1882 
1883        let's talk about how to access hoon that we write
1884 
1885        When we write hoon in unix, we can save it in the /gen folder of
1886        home, commit it , and then use it in dojo w/ + (lus) before the
1887        name of the hoon
1888 
1889        I don't know if you've read the Dojo getting started thing but
1890        that means effectively /home/gen/nameofgen/hoon
1891 
1892        in Urbit you don't use . (dot) to denote file extensions, it's
1893        just another fas (/)
1894 
1895        test
1896 
1897~rocted Ok I think I'm getting the hang of this BECAUSE I'm drawing the
1898        analogies to excel and it's finally making sense for me. I still
1899        cannot tell how the user input plays into this, but it sounds
1900        like your created program might give us an example of that. Feel
1901        free to teach more (this is great) or share the pastebin of that
1902        program (er...generator) and I'll check it out
1903 
1904        oh sorry. great!
1905 
1906~rabsef Are you comfortable w/ writing code to your unix environment for
1907        your moon, using vim?
1908 
1909~rocted yes, ok got it.
1910 
1911~rabsef I know vim /ok/ so that's what I use
1912 
1913~rocted not with vim but I use vscode. for now that is best fo rme
1914 
1915~rabsef have you mounted your moon?
1916 
1917~rocted for me
1918 
1919        yes sir!
1920 
1921        I have the /gen directory open in vscode
1922 
1923~rabsef nice - ok lets review what we want to do
1924 
1925        we want to create a gen that uses bartis |= to take in 2 atoms
1926        (@) and does add x y to them
1927 
1928        in what context would we normally see 2 atoms in hoon, recalling
1929        our original starting point
1930 
1931~rocted correct, with the atoms being submitted by the user correct?
1932 
1933~rabsef that's right -user submitted atoms
1934 
1935~rocted [2 4] - that context?
1936 
1937~rabsef well let's put that slightly differntly. bartis |= will /expect/
1938        values that we tell it to expect, so it could come from other
1939        hoon, too
1940 
1941~rocted in a cell?
1942 
1943~rabsef but yes - in this case user input
1944 
1945        ok - pastebin - one sec
1946 
1947~rocted when you say 'come from other hoon' - what is the "hoon" you
1948        mean here?
1949 
1950~rabsef https://pastebin.com/1r5mBUdW
1951 
1952        any hoon, suppose we fed values to our new program from another
1953        program
1954 
1955        hoon talks to hoon and contains hoon
1956 
1957        that is confusing but it's all reflexive to "the state"
1958 
1959        in the case of the program calling our new program, lets say,
1960        the input would not come from a user
1961 
1962        that's all I meant
1963 
1964~rocted ok i have copied that into vscode. going to rearrange my windows
1965        so I can see everything at once.
1966 
1967~rabsef one more rune we need to discuss ::
1968 
1969        :: allows comments
1970 
1971        also from a syntactic position, completely UNLIKE excel, hoon
1972        cares about spacing
1973 
1974        double space between rune and <contents>
1975 
1976~rocted copy that.
1977 
1978        thank you. did I read somewhere that double space and line feed
1979        are semantically equivalent?
1980 
1981~rabsef We can try that to see - I don't understand tall form sometimes
1982        tho because the spacing still matters
1983 
1984        I like long form
1985 
1986        oh big note here
1987 
1988        when you get errors in hoon that you write, you get [x y] which
1989        is always the row, column where the error happened
1990 
1991~rocted ok got it
1992 
1993~rabsef so if we fuck this up you'll be able to know
1994 
1995        where
1996 
1997~rocted interesting, and what is the column then? row is easy, but
1998        column?
1999 
2000~rabsef the # of chars to the right
2001 
2002~rocted (oh and also incidentally, I recall that someone said in u-h
2003        that sometimes it tells you where it discovered the error, but
2004        not where the error actually began).
2005 
2006        ok got it
2007 
2008~rabsef so for instance, bartis in my code is located at [8 1]
2009 
2010        yah that's right but u can trace it back
2011 
2012        so lets look at bartis and what it's saying there |= [a=@ b=@]
2013 
2014~rocted ah got it
2015 
2016~rabsef here a and b are arbitrary, they could be foo and bar, or crap
2017        and hat
2018 
2019        they're faces
2020 
2021        a "face" is a variable name
2022 
2023        or a constant name
2024 
2025        it's a name lol
2026 
2027~rocted ok cool.
2028 
2029~rabsef I like abcd because it's easy for me to remember
2030 
2031~rocted face. so in excel what would a 'face' analogue be?
2032 
2033~rabsef so - create a file in /gen that you call something other than
2034        add, and let's have just lines 8 and 14 in there
2035 
2036        a named range
2037 
2038        are you familiar w/ named ranges?
2039 
2040~rocted no unfortunately.
2041 
2042~rabsef or, specifically A1 would be teh face of the contents of that
2043        cell
2044 
2045        A1 is not the cell, nor the contents (except in weird cases) but
2046        it is the name of the cell
2047 
2048        Does that make sense?
2049 
2050~rocted not entirely because the binary tree provides addresses for each
2051        item, so those would be the best analogues to A1, B2
2052 
2053~rabsef A named range just allows you to give a cell or set of cells a
2054        funny name like "foo" or "a"
2055 
2056        well then go w/ the named range example - I have named the range
2057        of A1:A1 a
2058 
2059~rocted ah! ok.
2060 
2061~rabsef and B1:B1 b
2062 
2063~rocted copy that.
2064 
2065~rabsef Incidentally, those are very helpful in excel so you should try
2066        it out
2067 
2068~rocted oh can you actually do that in excel? Give a name to a range and
2069        then use that name over and over again? wow. could have saved me
2070        an hour of my life knowing that
2071 
2072~rabsef ok - so lines 8 and 14 in a file in gen called something.hoon -
2073        where something /isnt/ add because that already has a name
2074 
2075        and we can't unname add
2076 
2077~rocted got it
2078 
2079~rabsef yes, you can, and it does save a ton of time. You can even make
2080        them dynamic so they expand as the range gets more data, using
2081        offset function
2082 
2083        For instance, on the file name, my file is named testadd.hoon
2084 
2085        once I've saved that I need to commit home bar commit cen home
2086 
2087        or |commit %home
2088 
2089~rocted incidentally I also have the add item in the standard library
2090        also open
2091 
2092~rabsef You know, it would have been smart of me to look @ that
2093 
2094        lol
2095 
2096        when you commit home on your commet it will take a second and
2097        then it should show you that it read that file into the state
2098 
2099        lmk when done
2100 
2101        it will explicitly name your file in dojo
2102 
2103        ere moon
2104 
2105~rocted ok going to commit
2106 
2107~rabsef sorry to denegrate your moon
2108 
2109~rocted ok now how do we use this new +testadd program?
2110 
2111~rabsef in dojo but let's think about what we said we were going to give
2112        the program
2113 
2114~rocted commit complete (pretty quickly)
2115 
2116~rabsef I just spammed myself testing something that was wrong to see
2117        whta it would do
2118 
2119        knowing that we call a generator and knowing what our a=spec
2120        was, what do you think we should do
2121 
2122        (point of my spam message was i just cleared the screen so don't
2123        see anything above pretty quickly)
2124 
2125        also - let's talk about number format in hoon/urbit while you
2126        think about it
2127 
2128~rocted I think we should do +testadd [2 3] and it will output 5
2129 
2130~rabsef you're familiar w/ the number format? Like roman numerals,
2131        nothing greater than 3 numbers together
2132 
2133~rocted nope!
2134 
2135~rabsef ok - so 300 is fine
2136 
2137~rocted sorry nope was because I caused an error
2138 
2139~rabsef that should work actually lol
2140 
2141        one sec
2142 
2143        what do you get when you try that
2144 
2145        to be clear you did name your function testadd.hoon
2146 
2147        ?
2148 
2149~rocted let me check, I thought so
2150 
2151        yes, testadd.hoon, located in gen
2152 
2153        got an error at [20 68]
2154 
2155        I think it needs a final LF
2156 
2157        aha it worked!
2158 
2159        needed a final LF
2160 
2161~rabsef oh so you have the full thing w/ comments
2162 
2163        sorry - i just was thinking lines 8 and 14, meaning only 2 rows
2164        and Iw as like [20?~
2165 
2166~rocted ok great!!
2167 
2168~rabsef LF = carriage return?
2169 
2170        yah nice right? So let's make our function slightly more
2171        interesting. Let's add in a value unseen to the user that is
2172        also added
2173 
2174        let me write that really quickly and test it - probably 2 min
2175 
2176~rocted yes LF = carriage return but specifically the LF type, not the
2177        CRLF type
2178 
2179        ok cool. this is fun
2180 
2181~rabsef got it - one min
2182 
2183        ok - wrote it - going to commit but I'm concerned it might take
2184        a second
2185 
2186        so brb
2187 
2188        I have no eyebrows - somlus can you fix me?
2189 
2190        ok I wrote it and committed it and we're good
2191 
2192        We need to learn a new rune to do this
2193 
2194        tislas
2195 
2196        https://urbit.org/docs/reference/hoon-expressions/rune/tis/
2197 
2198        find tislas in there and look @ the first example of the syntax
2199 
2200~rocted reading...
2201 
2202        I'm glad we're doing this be cause this means of describing is
2203        beginning to get confusing for me.
2204 
2205~somlus there are a number of things I do not understand in this tis
2206        text:
2207 
2208        "some old subject p" - really? old subject from where? and
2209        what's a subject to be exact?
2210 
2211        "So to pin 12 with the face a" - where did the face a come from
2212        in this example?
2213 
2214~rabsef that first example looks a lot like what we're doing
2215 
2216        wouldn' tyou say
2217 
2218        which means - so I'm clear. Me?
2219 
2220        Read only "discussion" - that's all we need to know
2221 
2222        the rest of it is obnoxiously dense
2223 
2224        dotpost
2225 
2226        .
2227 
2228~somlus ok
2229 
2230        got a lot all at once, will read the discussion
2231 
2232~rabsef p, the value given to tislas can be of two types. It can have a
2233        type, or it can have no type
2234 
2235        recall =| a=@ <- this is typed
2236 
2237        ere |= rather
2238 
2239        tislas unlike bartis can get typed or not typed. Let's do typed
2240        for our sake
2241 
2242~somlus sorry sir there are a lot of discussions because there are a lot
2243        of uses of tis, so I'm a bit lost.
2244 
2245~rabsef note the syntax of the second line of the example under
2246        discussion
2247 
2248        Discussion
2249 
2250        p can be either a name or a name=type. If it's just a name, =/
2251        ("tisfas") "declares a type-inferred variable." If it has a
2252        type, =/ "declares a type-checked variable."
2253 
2254        I'm saying it wrong it's tisfas, sorry
2255 
2256~somlus ok will find that.
2257 
2258        np
2259 
2260~rabsef I'm trying to use the spoken names so I learn them
2261 
2262~somlus no problem. i'm getting a bit into a confused state so I'm going
2263        to take a bit and try to absorb this discussion section...
2264 
2265~rabsef well let's just do one thing before that
2266 
2267~somlus ok
2268 
2269~rabsef you don't need to know much but this - bartis lets user input
2270        values, tisfas declares a value in the hidden hoon
2271 
2272        we have bartis already taking in a cell. We want to add lets say
2273        2 to it covertly, behind the scenes
2274 
2275        new pastebin coming
2276 
2277        https://pastebin.com/jD5SabXA
2278 
2279        I've used 1 instead of 2 but it's the same difference. Find line
2280        14, I've given 1 a face of c
2281 
2282~somlus ok cool
2283 
2284~rabsef also my add syntax is redundant and I could have just (add a b
2285        c) but my way works too
2286 
2287        damned transitive property
2288 
2289        tbh i didn't realize add could take n values
2290 
2291~somlus ok I added that file. what will you call it so that we have same
2292        names? I'll save and commit
2293 
2294~rabsef I re-saved it as testadd
2295 
2296~somlus no worries we are learning together.
2297 
2298~rabsef since we don't need consistency or persistence of our original
2299        function
2300 
2301~somlus ok I'll call it testadd2.hoon because I might want to go through
2302        the entire chat again on my own.
2303 
2304~rabsef roger that - so if we resave as such (not sure how numbers work
2305        in the name incidentally, probably ok)
2306 
2307        if we +testadd2 [1 1] what should we get?
2308 
2309~somlus ok cool, got it, it adds 1 in addition to the items
2310 
2311        should be 3
2312 
2313~rabsef what do we get?
2314 
2315~somlus boom it worked
2316 
2317        3
2318 
2319~rabsef f'in a man
2320 
2321        right?
2322 
2323        pretty cool
2324 
2325~somlus yeah!!! mothef-er!
2326 
2327        awesome.
2328 
2329        ok so question
2330 
2331~rabsef let me pull up hoon school activity week 2 - I'm going to get
2332        that done this week and we can go thru that next
2333 
2334~somlus can I also say that =/ will set the value of a variable? or is
2335        that not specific enough in CS terminology
2336 
2337~rabsef I'm getting close to my termination point - need to eat dinner
2338        at some point
2339 
2340        hit me w/ ur question
2341 
2342~somlus yeah me too, but this is f-ing awesome. so fun.
2343 
2344~rabsef So I think technically =/ sets a constant
2345 
2346        right?
2347 
2348~somlus yeah ok.
2349 
2350        right
2351 
2352~rabsef because we're not changing 1 by the rest of our hoon
2353 
2354~somlus I think I should not burn my brain out but, in addition to the
2355        learning, we have also figured an awesome conversation flow and
2356        work flow to get this done. This is great!
2357 
2358~rabsef but yah a variable is just a changing constant or rather a
2359        constant is a trivial variable
2360 
2361~somlus correct, unless we rename it later with another constant, so yes
2362 
2363~rabsef Yah I really like this - I learn a lot by telling
2364 
2365~somlus I figured that might be a side benefit, but on my end this is
2366        just fantastic.
2367 
2368~rabsef trivial is a term they use a lot - it means the dumbest case of
2369        something
2370 
2371        a leaf is a trivial branch
2372 
2373        an atom is a trivial noun
2374 
2375~somlus ah got it that is helpful
2376 
2377~rabsef perhaps a commet is a trivial @p (name of ship aura)
2378 
2379~somlus makes sense, for sure.
2380 
2381        ok I think i actually have enough to go on here that I want
2382        to review the whole conversation and code again on my own. If
2383        you're ok with it we can call it a night, I'll do my review and
2384        we can pick up if you like at some convenient time that works
2385        for us down the road. This is really great!
2386 
2387        Thanks!
2388 
2389~rabsef one more thing - let me check that assignment and give a
2390        prospectus as to where we're going
2391 
2392        ok - next time we are going to make a hoon that checks if an
2393        incoming noun is a cell or atom, that then outputs that and then
2394        finally checks, if atom, whether it is odd or even
2395 
2396~tocref The assignment looks lovely!! I recommend giving constants and
2397        variables more descriptive names ususally, but in something like
2398        this it's a bit self-explanatory so it's not necessary
2399 
2400        The longer names are for your sake, as it's easier to figure out
2401        what everything means coming back to it after a week or so ––
2402        It's also a massive help to anyone else who reads your code
2403 
2404        This applies to every modern programming language, not just hoon
2405 
2406~somlus awesome rabsef, looking forward. I will solidify what we covered
2407        today so I am not deadweight as we move forward.
2408 
2409~rabsef that is a very good call toc
2410 
2411        Next week we will have complex names
2412 
2413        meaningful ones
2414 
2415        not dead weight at all - the questions draw attention to what I
2416        do not understand
2417 
2418        which is great compared to that which I do
2419 
2420~somlus understood tocref, so you are referring to the a, b, and c we
2421        used, and do you have any suggestions as to what cultural norms
2422        make sens for naming variables?
2423 
2424        ok great! so glad this has value. Had loads of fun by the way.
2425 
2426~rabsef likewise
2427 
2428        this was a great idea
2429 
2430~somlus And thanks for the outstanding comments on the two programs!
2431 
2432~tocref Here you're literally adding 3 numbers so it's hard to be more
2433        descriptive lol
2434 
2435~rabsef ok boys I got the dinner bell, and don't call me late
2436 
2437~tocref Especially because you also comment a line-by-line description
2438        of how it works
2439 
2440        num-1, num-2, and num-3 would work well (hoon styleguide likes
2441        all lowercase with dashes as spaces for variable names)
2442 
2443~rocted ok understood lowercase, dashes, and just a shortened version of
2444        exactly what it is. thanks.
2445 
2446~rabsef be on later boys - probably
2447 
2448        join ~rabsef-bicrym/lbsandkilos if you're into fitness - I feel
2449        I get a free plug for all that typing
2450 
2451        lol
2452 
2453~somlus ok will do!
2454 
2455        I'm going to close out for the night. thanks for all the help!
2456 
2457~bannum dot
2458 
2459~ronreg ~rasbef-bicrym: that's an interesting idea and I gave it a shot
2460        myself
2461 
2462        but I don't know why my solution isn't compiling
2463 
2464        ( https://pastebin.com/eW1bxbNR )
2465 
2466        I'm not sure if you were suggesting doing this in some kind of
2467        structured way
2468 
2469~somlus I say we do it whenever and ongoing. I've just been on a 10 hour
2470        flight so I was quiet
2471 
2472~rabsef ~ronreg-ribdev - I haven't finished yet so I don't want to look
2473        @ your solution but I will once I've finished coding mine
2474 
2475        terribly sorry
2476 
2477~risruc https://pastebin.com/tPRPkgrP
2478 
2479        here's what i got
2480 
2481~filsur /leave ~paldev/nopub
2482 
2483        chats;
2484 
2485~ronreg ~risruc-habteb: looking at your solution, I think my problem was
2486        something involving types
2487 
2488        I was trying to use %odd-atom %even-atom and %cell as the thing
2489        I returned, but I suspect that doesn't actually type-check
2490 
2491~risruc I'm not sure what is the proper way to do what you were trying.
2492 
2493        above my pay grade!
2494 
2495~happyl test
2496 
2497~fabled test works
2498 
2499~ronreg ~risruc-habteb: your solution involves returning one of three
2500        strings to indicate which of the three possibilities the input
2501        was, whereas I was trying to do this by returning one of three
2502        %symbols
2503 
2504        but I think that might be a type error, to have multiple
2505        branches of a conditional return different %symbols
2506 
2507        on the other hand, `?: =(1 1) %hella %brah` works in the dojo
2508        (it returns %hella)
2509 
2510        so maybe I'm still misunderstanding some finer point about
2511        %symbols :)
2512 
2513~rabsef I'm ready to do next week
2514 
2515        https://pastebin.com/1rnWp8nt
2516 
2517        Let's (~somlus when you're ready and back stateside?) plan
2518        another sesh - I really enjoyed that
2519 
2520        ~ronreg - this fixes your solution. You weren't using wuttis
2521        correctly (I think that takes p type q hoon)
2522 
2523        https://pastebin.com/SY3GNipp
2524 
2525        but also that's not technically a naked gen, that's a %say
2526        generator
2527 
2528~ronreg ah okay
2529 
2530        ah okay so ?= was just totally the wrong rune to use
2531 
2532        so it was a type error, not just the one I thought it was :)
2533 
2534~rabsef you can use ?= in a dofferent context tho as the head of wutcol
2535 
2536        but yah :)
2537 
2538~somlus totally I'm in for another sesh for sure next week or if I can
2539        work out my schedule over here even sooner
2540 
2541~rabsef Assignment 3 today - I did fibonacci sequence previously so this
2542        one should be pretty easy, but the tail recursion is still a
2543        little confusing
2544 
2545        so I did it but I can’t get the error checking right
2546 
2547        sick - that's at least 2 pieces of content we can work thru
2548 
2549        Got it
2550 
2551~somlus I'm impressed with your persistence rabsef! Will dive back in,
2552        in the next few days and keep you posted.
2553 
2554~rabsef I’m going to make something in this before I die. I will, with
2555        my will.
2556 
2557        maybe I should say explicit children - they can be discrete but
2558        needn’t be
2559 
2560        the best part of Hoon is discrete children of operations,
2561        meaning you never get an excel “you forgot a close paren
2562        somewhere” thing
2563 
2564~ronreg yeah the syntax of hoon is very interesting
2565 
2566~rabsef speaking of, ~somlus - you should check out the 1.5 Lists
2567        walkthrough and familiarize yourself w/ the basic concepts, as
2568        well as 1.4.1 Recursion - we can go through both of those and do
2569        all the exercises together for next session
2570 
2571~somlus Trying to work out the international time zone thing. It looks
2572        like unless you keep odd hours, the only reasonable period to do
2573        this is your morning, my later afternoon. Was hoping to get some
2574        things down on my own though before our next hoontards sesh.
2575        Would your preference to be to charge ahead tomorrow (Sunday),
2576        or New Years' eve or New Years' day (morning) or just wait a
2577        week or so until I return to the states?
2578 
2579        ~rabsef
2580 
2581~rabsef I am happy to do it when you feel ready. It sounds like that
2582        would be maybe when you return? Giving you time to forge ahead
2583        and gather questions - absolutely no rush
2584 
2585        lol
2586 
2587        just wrote fibonnaci in 1 try - I feel learning happening
2588 
2589        fibonacci*
2590 
2591~socsen Hi! I'm not officially apart of hoon school but I've been going
2592        through the documentation on my own...... Is there a link to the
2593        assignments? Would like to give them a shot.
2594 
2595~rabsef ~socsen - join ~borlyt-salnus/file-exchange
2596 
2597        I posted them there for someone else
2598 
2599~socsen I tried joining but got: permission-hook failed subscribe on
2600        /chat/~borlyt-salnus/file-exchange/read
2601 
2602~rabsef I've asked the owner of that if they'll invite you
2603 
2604        I can post it here later, or make another forum, but I don't
2605        want to spam if I can avoid
2606 
2607        Apropos of hoon school, I just finished lesson 4 - 2 more and 6
2608        hasn't been posted yet. I'm going to catch up this weekend.
2609 
2610~socsen Thank you :)
2611 
2612~somlus Testing a post from beneath the english channel on a bullet
2613        train.
2614 
2615        Testing a post from beneath the english channel on a bullet
2616        train.
2617 
2618        Testing a post from beneath the english channel on a bullet
2619        train.
2620 
2621        Testing a post from beneath the english channel on a bullet
2622        train.
2623 
2624        Testing a post from beneath the english channel on a bullet
2625        train.
2626 
2627        Testing a post from beneath the english channel on a bullet
2628        train.
2629 
2630        Testing a post from beneath the english channel on a bullet
2631        train.
2632 
2633        Testing a post from beneath the english channel on a bullet
2634        train.
2635 
2636        Testing a post from beneath the english channel on a bullet
2637        train.
2638 
2639        Testing a post from beneath the english channel on a bullet
2640        train.
2641 
2642        Testing a post from beneath the english channel on a bullet
2643        train.
2644 
2645        Testing a post from beneath the english channel on a bullet
2646        train.
2647 
2648        Testing a post from beneath the english channel on a bullet
2649        train.
2650 
2651        Testing a post from beneath the english channel on a bullet
2652        train.
2653 
2654        Testing a post from beneath the english channel on a bullet
2655        train.
2656 
2657        Testing a post from beneath the english channel on a bullet
2658        train.
2659 
2660        Testing a post from beneath the english channel on a bullet
2661        train.
2662 
2663        Testing a post from beneath the english channel on a bullet
2664        train.
2665 
2666        Testing a post from beneath the english channel on a bullet
2667        train.
2668 
2669        Testing a post from beneath the english channel on a bullet
2670        train.
2671 
2672        Testing a post from beneath the english channel on a bullet
2673        train.
2674 
2675        Testing a post from beneath the english channel on a bullet
2676        train.
2677 
2678        Testing a post from beneath the english channel on a bullet
2679        train.
2680 
2681        Testing a post from beneath the english channel on a bullet
2682        train.
2683 
2684        Testing a post from beneath the english channel on a bullet
2685        train.
2686 
2687        Testing a post from beneath the english channel on a bullet
2688        train.
2689 
2690        Testing a post from beneath the english channel on a bullet
2691        train.
2692 
2693        Testing a post from beneath the english channel on a bullet
2694        train.
2695 
2696        Testing a post from beneath the english channel on a bullet
2697        train.
2698 
2699        Testing a post from beneath the english channel on a bullet
2700        train.
2701 
2702        Testing a post from beneath the english channel on a bullet
2703        train.
2704 
2705        Testing a post from beneath the english channel on a bullet
2706        train.
2707 
2708        Testing a post from beneath the english channel on a bullet
2709        train.
2710 
2711        Testing a post from beneath the english channel on a bullet
2712        train.
2713 
2714        Testing a post from beneath the english channel on a bullet
2715        train.
2716 
2717        Testing a post from beneath the english channel on a bullet
2718        train.
2719 
2720        Testing a post from beneath the english channel on a bullet
2721        train.
2722 
2723        Testing a post from beneath the english channel on a bullet
2724        train.
2725 
2726        Testing a post from beneath the english channel on a bullet
2727        train.
2728 
2729        Testing a post from beneath the english channel on a bullet
2730        train.
2731 
2732        Testing a post from beneath the english channel on a bullet
2733        train.
2734 
2735        Testing a post from beneath the english channel on a bullet
2736        train.
2737 
2738        Testing a post from beneath the english channel on a bullet
2739        train.
2740 
2741        Testing a post from beneath the english channel on a bullet
2742        train.
2743 
2744        Testing a post from beneath the english channel on a bullet
2745        train.
2746 
2747        Testing a post from beneath the english channel on a bullet
2748        train.
2749 
2750~rabsef lol received
2751 
2752        also that’s awesome
2753 
2754~rivmep spam test
2755 
2756~somlus wow! I wonder why it spikes a ton of echoes like that when I'm
2757        on a weak signal. Well, there you have it. at least it worked!
2758 
2759        If anyone was wondering, that was a 46 item echo from a single
2760        post.
2761 
2762~rabsef How was the train, m8?
2763 
2764~somlus wow we humans are badasses. What wonderful technology that was
2765        to zip along, comfortably swaying at a fast clip, and then dip
2766        beneath the ocean (holy cow) and pop up in Great Britain.
2767 
2768~rabsef how long and did you have a sleeper/couchette?
2769 
2770        I remember taking train trips with my grandfather and getting to
2771        sleep in the like drop down bed from a bulkhead in the sleeper
2772 
2773~hanryc lol
2774 
2775~rabsef ~socsen - you should have gotten an invite to file-exchange
2776 
2777        ~borlyt (owner) said they would get it to you
2778 
2779~somlus Sorry for the delay. No sleeper needed! Paris to London in just
2780        over 3 hours. The scenery was interesting and fast enough to
2781        keep us captivated. We did travel in premium seats which came
2782        with a pretty good meal!
2783 
2784~rabsef nice - rail travel is p. dope
2785 
2786~ronreg one of the things I'm finding in the process of trying to learn
2787        hoon
2788 
2789        is that I have trouble keeping straight lots of concepts
2790        pertaining to cores
2791 
2792        exactly what a battery, sample, payload, etc. are
2793 
2794        and what the semantics of core arms/legs are wrt the subject
2795 
2796        I can conceptualize a core as a collection of functions that are
2797        either grouped together for organizational reasons or because
2798        they operate on shared data - that is pretty similar to a
2799        classic OOP class
2800 
2801~rabsef I think that stuff is complex too. battery = operations and
2802        payload = sample + context, which is the totality of the subject
2803        as far as I can tell. except, the core itself is either part of
2804        the context or sample and therefore seemingly the battery is in
2805        the payload
2806 
2807        lol
2808 
2809~socsen ~rabsef I believe I've joined but the chat is not showing up on
2810        my chats menu and I'm not receiving any messages :(
2811 
2812~rabsef ~socsen - does it show in ;what from chat-cli?
2813 
2814        I assume this means you're still pulling the backlog and it's
2815        taking a while
2816 
2817        this is normal - and may take a little longer
2818 
2819~somlus How is it now worsen?
2820 
2821        Sorry autocorrect. meant socsen
2822 
2823~socsen Still no luck. My chat cli is
2824        :chat-cli[~borlyt-salnus/file-exchange]
2825 
2826        I see what you're asking....... after typing ;what that chat
2827        isn't showing up
2828 
2829        I've inspected the network calls and the array that comes back
2830        says I have permissions to join so I'm not sure.
2831 
2832~rabsef You appear to be in it
2833 
2834        try binding it - I think it's ;bind glyph chat
2835 
2836        also try |hi-ing the host ~borlyt-salnus
2837 
2838        Also, I did just repost it there so when you do get access, you
2839        should see it.
2840 
2841~socsen Still no luck grrr..... I tried |hi-ing the host ~borlyt-salnus
2842        that came back succesful.......... i also tried |hi-ing ~borlyt
2843        but this just hangs for me
2844 
2845~rabsef https://cdn.discordapp.com/attachments/502183430583484418/661239
2846        661863108628/Welcome_to_Hoon_School_101.docx
2847 
2848        ~borlyt was kind enough to provide this
2849 
2850        https://cdn.discordapp.com/attachments/502183430583484418/661239
2851        666279841812/Welcome_to_Hoon_School_101.pdf
2852 
2853~socsen <3
2854 
2855        Thank you I'm sorry this has been so much trouble
2856 
2857~rabsef do ;bind <a glyph you haven't used yet in ;what> ~borlyt-salnus
2858        in chat-cli
2859 
2860~somlus Hey were you doing that so you could get the hoon school
2861        assignments? I forgot the purpose
2862 
2863        Here are two other versions that might help:
2864 
2865        https://www.dropbox.com/s/bre7oib2qasd519/Hoon%20101%20Course%20
2866        Materials.pdf?dl=0
2867 
2868        https://www.dropbox.com/s/d4vguf42q17t6om/Hoon%20101%20school%20
2869        posts%20%28updated%20version%20with%20links%29.docx?dl=0
2870 
2871        Hey question guys - when I posted "Testing a post from beneath
2872        the english channel on a bullet train." - did you also see it
2873        like 40+ times in a row or was that just on my end? Submitted a
2874        github issue and they asked me.
2875 
2876~socsen Yes that was the purpose! Thank you!
2877 
2878        I saw it like 40 times^^
2879 
2880~rabsef likewise
2881 
2882~somlus thanks for that info.
2883 
2884        socsen I updated the word document file with week 6 stuff so if
2885        you need that one please download it again. Should be the same
2886        link as the filename is the same.
2887 
2888~ronreg hoon exercise that might be interesting
2889 
2890        (that I just did)
2891 
2892        take the code in :
2893        https://urbit.org/docs/tutorials/hoon/libraries/
2894 
2895        and edit it such that the value of a playing card is an actual
2896        playing card value (ace, 2, 5, king, etc.) rather than just an
2897        integer
2898 
2899~minder ~ronreg, this is the content I'm here for.
2900 
2901~somlus Sir rabsef! How's it going today, and how are your synapses
2902        reorganizing around hoon these days?
2903 
2904        Just read in Minder's party that you did week 5 of hoon school -
2905        impressive.
2906 
2907~rabsef I am beginning to awaken
2908 
2909        i feel it percolating up from the subject of my mind to the
2910        operational arms as samples and context
2911 
2912        i would call myself a beginner now, after being a nothing
2913 
2914        I am excited to do some more with you.
2915 
2916~somlus That is truly inspiring. I love the moment in learning when
2917        the head fog (and insecurity) starts to clear. Slowly becomes
2918        focus/understanding/power, which rocks.
2919 
2920~rabsef I will get there - I will not give up on this
2921 
2922        I'm going to make something, so help me lol
2923 
2924~somlus How's your Sunday ~rabsef? I could do some hoontarding later
2925        this morning or later afternoon (PST).
2926 
2927~rabsef we can do after church, around 9:30 mountain so 8:30 your time
2928 
2929        i need to write first edition of 52 Sundays so I’ll be taking
2930        notes
2931 
2932        Build a naked generator that takes a noun and checks if that
2933        noun is a cell or an atom. If that input noun is an atom, check
2934        if it’s even or odd. The output should be of the tape type. A
2935        tape is a string.
2936 
2937        https://urbit.org/docs/tutorials/hoon/hoon-syntax/
2938 
2939        https://urbit.org/docs/tutorials/hoon/conditionals/
2940 
2941        lets look at this assignment and these articles
2942 
2943        we didn’t do that yet right?
2944 
2945~somlus Cool. I have gym time at 10am so that's a good hard stop right
2946        before that.
2947 
2948~rabsef heading back mow
2949 
2950~somlus ?
2951 
2952~fogdeg Ready as moon
2953 
2954        (i think)
2955 
2956        I booted this moon and it's taking a long time to sync the chat,
2957        but I can see that these messages are sending so we're pretty
2958        good.
2959 
2960        test
2961 
2962~rabsef hey I'm here
2963 
2964        sorry about taking so long
2965 
2966        Hey ~somlus-savlev - I am here
2967 
2968~fogdeg no worries!
2969 
2970        Just setting up my wife's ship again, had to breach
2971 
2972        how are you? How was the holiday? Looks like you remained fired
2973        up and excited to work on hoon.
2974 
2975~rabsef Nice. And yah - I actually got a lot of the work done. I'm on
2976        the last week and if I complete it today (iffy) then I will
2977        actually be submitting it on time, I think.
2978 
2979~fogdeg unbelievable! That's quite an achievement.
2980 
2981~rabsef I feel more comfortable than I have before in hoon and the
2982        assignments are coming w/ less and less help
2983 
2984~fogdeg Since we only have an hour or so, maybe it would be a good use
2985        of time to cover some "wish you had known" things - what do you
2986        think?
2987 
2988        I'm sure you have had lots of epiphanies in the last few weeks.
2989 
2990~rabsef Yes - that sounds good. I think the first thing to note is that
2991        there's a difference between a gate and a door. And, generally,
2992        what a gate is
2993 
2994        Are you familiar w/ gates at all so far?
2995 
2996        we made a gate last time when we asked for a noun (an argument)
2997        and cast it as a different type of noun
2998 
2999~fogdeg correct.
3000 
3001        What I should do is take more notes and review them before we
3002        meet again. I'll start doing that.
3003 
3004~rabsef Well I could have been more communicative about a plan but I
3005        think just following hoon 101 for now is probably pretty smart
3006 
3007~fogdeg oh that's great. I'll reinvigorate my verve to do that.
3008 
3009~rabsef So a gate and a door. A gate is a "core". A core is a battery
3010        payload cell [battery payload]. The battery an "arm" (some cores
3011        can have multiple arms using barcen, we'll get to that later)
3012 
3013        An arm is effectively an operation as part of a function. To
3014        review the "mathematical" definition of a function, a function
3015        is an operation where, given the same arguments, you get the
3016        same result, always
3017 
3018        this is big in hoon and urbit - it's why we have a deterministic
3019        system and why they say urbit and hoon are referentially
3020        transparent
3021 
3022        all operations, giv en the same steps, lead to the same rome as
3023        it were
3024 
3025~fogdeg yes. As you might recall I often find Excel comparisons helpful
3026        so if that applies here, feel free to make those analogies.
3027 
3028        Will read the above carefully.
3029 
3030~rabsef so back to gate qua core. A gate is a core w/ an arm (battery)
3031        and a payload. A payload is another cell [sample context]
3032 
3033        Ah yes in excel, a gate is effectively like this [=excel
3034        function [arguments to excel function - the whole rest of your
3035        worksheet]
3036 
3037        ^]
3038 
3039~fogdeg Battery is something that is done to something?
3040 
3041~rabsef yes
3042 
3043~fogdeg got it, i'm back on track now
3044 
3045        payload is like a type of output but in the format of a cell?
3046 
3047~rabsef [Excel Function [Arguments - Worksheet]] to clarify my typing
3048        there. So "worksheet" - this is why they call this a subject
3049        oriented programming language - this is something a little
3050        beyond my comprehension but basically, the entirety of the
3051        operating environment is always considered in your operations -
3052        if it is not necessary it is ignored but it is there to be used
3053        if you want
3054 
3055        payload is the arguments - so if we have say =sum, right? We
3056        might say the core would be like [=sum [(1, 2) (everything else
3057        on the worksheet)]]
3058 
3059        we would get 3 from this because sum only uses "the sample"
3060 
3061~fogdeg why even include the (everything else in the worksheet)?
3062 
3063        if it is to be ignored?
3064 
3065~rabsef because it often is not ignored. In this way you needn't
3066        reprogram (mul) the hoon multiply program in every program you
3067        write
3068 
3069        the standard library of programs available to you , and indeed
3070        any programs you write become part of the context
3071 
3072        [battery [sample context]]
3073 
3074~fogdeg ok got it
3075 
3076~rabsef ok so when you create a gate, you need to provide an argument
3077        (mostly) or else it's a door (I think). Doors can trip w/o
3078        samples. But let's focus on gates.
3079 
3080        And this will make context more clear
3081 
3082~fogdeg ok
3083 
3084        perfect
3085 
3086~rabsef You can do this in dojo. You can create a bound value in the
3087        context "=b 10"
3088 
3089~fogdeg done
3090 
3091~rabsef now if you create a gate, you can use b indiscriminately, even
3092        if you unbind b
3093 
3094        so lets make that gate as this "=tenlus |=(a=@ (add a b))"
3095 
3096        what should that do?
3097 
3098        if we operate it
3099 
3100        Incidentally, I think it is called kicking a gate, and slamming
3101        a door when you operate them
3102 
3103        Let's go throug heach part. =tenlus just gives us the name of
3104        the function
3105 
3106~fogdeg ok
3107 
3108~rabsef |= is how you call the gate and a=@ tells us we have what sort
3109        of argument?
3110 
3111~fogdeg got it
3112 
3113~rabsef or rather that tenlus expects what sort of sample?
3114 
3115~fogdeg understood, since we used that before
3116 
3117~rabsef right so that's a noun that is an atom in this case, with no
3118        further context (typing)
3119 
3120        then add is fairly clear, so =tenlus will give us a + b where a
3121        is our sample and b comes from the context
3122 
3123        [battery [sample context]]]
3124 
3125        let's try it (tenlus 15) should result in 25
3126 
3127~fogdeg got it, so in plain english you created a function to add 10 to
3128        a real number that is input in the dojo
3129 
3130        bingo
3131 
3132~rabsef technically we're not saying a real number I guess - we're
3133        saying an atom which could be an @p or any number of types, save
3134        a cell or a tape or other types that are non-atomic
3135 
3136        but yes - otherwise. Just bear in mind that it could be
3137        something strange and non numeric
3138 
3139        the reason for that is that all values in hoon are effectively
3140        numbers in nock
3141 
3142        if you were to `@ud`~somlus-savlev in dojo you will get your
3143        planet number
3144 
3145~fogdeg got it. In a way I'm still "translating" to my pre-hoon way of
3146        describing things. That's why I do that. I know that eventually
3147        I'll need to use this new vocabulary because it is more precise
3148        for this environment
3149 
3150        yes, got it.
3151 
3152~rabsef I think it's important to make referential comparisons tho -
3153        believe me when I say I'm trying that too lol
3154 
3155~fogdeg Agreed - it's a temporary crutch that we both hope to graduate
3156        from.
3157 
3158~rabsef ok - so important note. You can not only modify your sample
3159        (e.g. (tenlus 25)) but you can modify your context - they call
3160        this making a mutant version of teh core w/ a mutant version of
3161        the context
3162 
3163~fogdeg can you explain context further?
3164 
3165~rabsef oh first - let's prove that the context is bound into the core
3166 
3167~fogdeg ok
3168 
3169~rabsef unbind b 1) first hit b in dojo and enter, and you should get 10
3170 
3171        then 2) =b with nothing following and enter unbinds it
3172 
3173        3) b again should give you a find.b error
3174 
3175        now try (tenlus 10) and you will still get 20
3176 
3177        this is also part of something important about hoon - I don't
3178        quite understand it but basically dependencies get rolled into
3179        cores by ford when they are compiled using |commit (or just as
3180        part of dojo actions which ford also manages)
3181 
3182~fogdeg interesting.
3183 
3184        so the value stored in b survives the existence ofb in a way
3185 
3186~rabsef the entirety of the "state" is rolled in, as needed, yes.
3187 
3188        "state" is something like operating environment at the time of
3189        creation, e.g. worksheet
3190 
3191~fogdeg perhaps it's a bit like 'commit' in that it is stored in two
3192        places and changing it in one does not necessarily change the
3193        other
3194 
3195        got it
3196 
3197        at the time of creation...
3198 
3199~rabsef now let's mutate tenlus
3200 
3201~fogdeg ok
3202 
3203~rabsef (tenlus(b 15) 10)
3204 
3205        will give us what
3206 
3207~fogdeg 25
3208 
3209        because b is now zero?
3210 
3211~rabsef and to restate what it's doing when it does that is recreating
3212        the core at runtime w/ a fake context. B is now 15 using the (b
3213        15) part of the above
3214 
3215        25 is right but b is either 10 in the normally defined tenlus,
3216        despite being unbound, or 15 in the mutant variet
3217 
3218        y*
3219 
3220~fogdeg ah I thought it was adding b and 15 and then adding 10, to get
3221        25
3222 
3223        it was defining b as 15
3224 
3225~rabsef correct - to the latter
3226 
3227        ok - what next lol
3228 
3229~fogdeg I do not fully understand "mutant variety" - is it not confusing
3230        to have two states to one thing?
3231 
3232~rabsef ok - good question. tenlus will forevermore (until we unbind it
3233        and rewrite it completely) have a state where b=10
3234 
3235        it is so written, it will be so done - hoon is just like that
3236 
3237~fogdeg ah! thank you for explaining that
3238 
3239~rabsef "mutating" - imagine you have a program that you want to test
3240        new values against - this would be the mutant case
3241 
3242~fogdeg got it.
3243 
3244~rabsef "Ford" is a vane that does that work for us - when it compiles
3245        a program of hoon, it takes the context into consideration,
3246        looks @ the relevant parts, and binds them into the program
3247        permanently
3248 
3249~fogdeg ok
3250 
3251~rabsef I /think/ that's actually a tiny bit misleading. Ford can also
3252        do something where, if the context changes (in a non mutant way)
3253        e.g. if you rebind b to something else now, it will know to
3254        recompile tenlus, but that's a little beyond our means right now
3255 
3256        Ted B. (~rovnys) did a good talk about leafs and recompiling but
3257        I don't really understand all of it
3258 
3259~fogdeg very interesting. Is there a 'big picture' way to describe why
3260        it is so important to do it this way? I am guessing because this
3261        system lives in a networked world that changes over time, so
3262        it's important to be able to calculate what was true in a prior
3263        time...
3264 
3265~rabsef it is enough to know, now, that a program uses the bound state
3266        at time of compile and that you can poke values in the context
3267 
3268~fogdeg ok I'll take a look at that talk. It's a link in the hoon school
3269        assignments?
3270 
3271        ok got it
3272 
3273~rabsef that's exactly it - if the context is important and if
3274        ford knows the context, it can recompile efficiently when
3275        dependedencies change
3276 
3277        e.g. when =b is rebound to lets say 15. In larger programs let's
3278        say chat-cli is dependent on ames, and ames changes, chat-cli
3279        will be recompiled by ford automatically
3280 
3281        this is how OTAs are done
3282 
3283        I'll find that talk and post it here.
3284 
3285~fogdeg ok that is helpful. how interesting and complex to create
3286        something like that.
3287 
3288~rabsef right?!
3289 
3290        lol
3291 
3292~fogdeg amazing.
3293 
3294~rabsef for the sake of brevity I'm going to give us the program for the
3295        assignment and we can quickly talk throug hit
3296 
3297~fogdeg cool
3298 
3299~rabsef The assignment again was to make a "naked generator" (i.e. a
3300        core w/ no special properties) that tells us if an incoming
3301        values is a cell, or an atom and if an atom, if odd or even
3302 
3303        as follows
3304 
3305        |= a=*
3306 
3307        ?: ?=(@ a)
3308 
3309        ?: =(0 (mod a 2))
3310 
3311        "atom-even"
3312 
3313        "atom-odd"
3314 
3315        "cell"
3316 
3317        there are several important things going on here
3318 
3319~fogdeg I'm following.
3320 
3321~rabsef First, remember that wutcol, the conditional (basic) always
3322        takes two children
3323 
3324        that's why we don't need to close paren around arguments as
3325        withexcel
3326 
3327~fogdeg got it
3328 
3329~rabsef ok so from top to bottom - we ask to form a gate, where the
3330        argument is * (a noun of any kind)
3331 
3332        second, we ask "is this noun an atom?" If so, we do the next
3333        3 lines, if not we output "cell" because there are only two
3334        options
3335 
3336        atom or cell
3337 
3338        we know we jump down to cell because, again, wutcol takes 2
3339        arguments (and for the second wutcol they are "atom-even" and
3340        "atom-odd")
3341 
3342        not arguments - something like conditional outputs
3343 
3344        one for %.y and one for %.n
3345 
3346        %.y and %.n are the booleans for yes and no, true and false,
3347        just as in excel
3348 
3349        they can also be abbrv. & for %.y and | for %.n
3350 
3351        are you familiar w/ the math operation modulo?
3352 
3353~fogdeg I am not
3354 
3355~rabsef I was not either. Modulo is important in programming, I guess,
3356        but it just means give me the remainder
3357 
3358        3 mod 2 is 1
3359 
3360        it goes in 1 time with 1 remaining
3361 
3362~fogdeg got it ok
3363 
3364~rabsef thus 5 mod 2 is also 1
3365 
3366~fogdeg got it ok
3367 
3368~rabsef so our second conditional says IF (a - our atom - mod 2)=0, then
3369        %.y, else %.n
3370 
3371        and wutocol ?: says if yes, do my first output if no do my
3372        second
3373 
3374        thus ?: =(0 (mod a 2)) "atom-even" "atom-odd" says what
3375 
3376~fogdeg ah!
3377 
3378~rabsef it becomes so clear as you learn it but it is hard to learn imo
3379        lol
3380 
3381~fogdeg so if there is no remainder, then of course it's even. if there
3382        is a remainder then it's odd
3383 
3384        yeah i am feeling that. Part of my goal is to continue to
3385        work on this so that I can eventually explain it very well to
3386        non-coders.
3387 
3388~rabsef Thus our formula in excel would be similar to this {bear in mind
3389        that there is no comparison to cell and atom in excel, really}
3390        =IF(a="atom",IF((a mod 2)=0,"even","odd"),"cell")
3391 
3392~fogdeg wow that's cool. and you're right we'd have t come up with a way
3393        of defining a cell in excel for that to work.
3394 
3395~rabsef we could say something like isnum or something
3396 
3397~fogdeg That excel reference actually really helps me.
3398 
3399~rabsef but that's still not right really - but for our purposes it's
3400        fine
3401 
3402        we must just remember that an atom is a leaf in a tree and a
3403        cell is a complex branch
3404 
3405~fogdeg yeah good point.
3406 
3407~rabsef an atom is a trivial branch of a tree, a leaf
3408 
3409~fogdeg that much I've got! thankfully
3410 
3411~rabsef yes great. Where to next
3412 
3413        let me think
3414 
3415~fogdeg good question. Maybe I should focus on this one example for a
3416        while. I have to leave for the gym in 15 mins anyway.
3417 
3418~rabsef yes - good - do you want to guide me through any parts you're
3419        having issue w/?
3420 
3421~fogdeg This is actually a very good exercise to learn more basics and
3422        also some logic language.
3423 
3424~rabsef It's extremely helpful to me to restate it
3425 
3426        lol
3427 
3428~fogdeg yes
3429 
3430        ok
3431 
3432~rabsef It's like studying only not boring
3433 
3434~fogdeg so ?: has two children but ?= is a simple question that
3435        immediately follows?
3436 
3437        that immediately follows ?=
3438 
3439        would that be correct?
3440 
3441~rabsef Yes - we might say ?: checks a condition and then outputs based
3442        on the boolean output of the condition first %.y and then %.n,
3443        while wuttis (?=) gives us either %.y or %.n
3444 
3445~fogdeg but either way there is a 'yes' and a 'no' option that follows
3446        right after that?
3447 
3448~rabsef Not exactly. Some wut commands give outputs that are just
3449        boolean. So ?= just outputs either the boolean true or the
3450        boolean false - it is unlike excel's if
3451 
3452        it can never give us something other than %.y or %.n
3453 
3454        wutcol on the other hand reads the boolean output of a condition
3455        (like wuttis) and then does operations based on that - this is
3456        far more like excel's if
3457 
3458        wuttis is like =IF(condition, TRUE, FALSE)
3459 
3460        and you could imagine =IF(IF(condition, TRUE, FALSE), True
3461        output, False output)
3462 
3463        see the difference there?
3464 
3465~fogdeg you mean wutcol or wuttis for that last comparison?
3466 
3467        oh yeah, ignore that question
3468 
3469~rabsef =IF{wutcol}(IF{wuttis}(condition, TRUE, FALSE)
3470 
3471        the important part is that the outputs of wuttis can not be
3472        changed to be other hoon or other outputs like "atom-even"
3473 
3474~fogdeg got it.
3475 
3476~rabsef while wutcol can be made to do more hoon
3477 
3478~fogdeg ok awesome, I have that now.
3479 
3480~rabsef Here is one more thing we will need soon - the rune kethep ^-
3481 
3482        kethep can be used in a gate to define the output type
3483 
3484        in our case we could have ^- tape
3485 
3486        because "atom-even" "atom-odd" and "cell" are tapes
3487 
3488~fogdeg However in your code I do not see that you used either %.y or
3489        %.n - it seemed to just follow the code, which was to produce an
3490        output.
3491 
3492~rabsef but since we "" around them, they are already typed
3493 
3494~fogdeg ok got it. I will look into kethep too then
3495 
3496~rabsef consider this - ?: (hoon that results in either a %.y or %.n),
3497        %.y condition output, %.n condition output
3498 
3499        we could have done other hoon in the parenthetical insertion
3500        outside of ?= but we did something easier this time
3501 
3502~fogdeg yes but we used ?= and it theoretically would have just
3503        outputted %.y, and then presumably you would have used that %.y
3504        to output some text on condition %.y
3505 
3506~rabsef for instance, we could have done 'not (mod a 2) = 0)' and
3507        switched our output lines
3508 
3509~fogdeg but it went ahead and outputted the text so it seems that is
3510        more like ?:
3511 
3512~rabsef that is exactly what we did - your first assertion. The ?: uses
3513        the %.y to trigger "atom-even" and the %.n (of ?=) to trigger
3514        "atom-odd"
3515 
3516        ?= has no children - if we just had ?= there we would get a
3517        "nest fail" because the lines "atom-even" and "atom-odd" would
3518        not accurately nest under the other hoon present
3519 
3520        If we did that - if we just had ?=, the program would read - if
3521        a is an atom, then give me a yes or no as to whether a modulo 2
3522        is 0, else print "atom-even"
3523 
3524~fogdeg ok. I'll spend some more time making sure i have this down. It's
3525        nice to have it all here. I have to bounce to the gym (scheduled
3526        gym time) but will check in later for sure.
3527 
3528~rabsef and then "atom odd" and "cell" would be errors
3529 
3530~fogdeg ok I will look close a that soon to be sure I understand it.
3531 
3532~rabsef absolutely - let mek now if you have questions after thinking on
3533        it - I'l lbe online a lot today, ideally
3534 
3535~fogdeg Thanks again - this is very fun. Ok cooo, will do.
3536 
3537        cool
3538 
3539~rabsef have fun at gym :)
3540 
3541        ttyl
3542 
3543        I had something backwards when I was talking earlier. A gate
3544        is a special case of a door - I thought they were potentially
3545        distinct but I 've just learned and been corrected by the text.
3546        To clarify, a CORE is [battery payload], a DOOR is a core where
3547        the [payload] is [sample context], and a GATE is a door where
3548        [battery] is one arm named $, and [payload] is [sample context].
3549        that is to say a gate is a very simple door which could have
3550        multiple arms (like the %math standard library door) and each of
3551        the arms could potentially be a gate
3552 
3553        And, battery and payload are always battery -> the operation(s)
3554        to be done, payload -> the data required to complete those
3555        actions
3556 
3557        so potentially there are some cores that are not doors nor
3558        gates, but instead do not require a sample, or do not require
3559        the context
3560 
3561~somlus Thanks for this follow-up info - the chat did not autoupdate
3562        until i refeshed. I will take this into account when I dig
3563        deeper into those particular gates / arms. I am at the time of
3564        day when jet lag is the worst and I feel cognitively like I'm at
3565        the very end of an evening of drinking. Hopefully in a few hours
3566        / tomorrow I'll be back to sharp again! I guess it's good that
3567        we did our session in the morning, since I'm pretty useless at
3568        this very moment.
3569 
3570~ronreg I'm not clear on the distinction between cores and doors
3571 
3572        I'm still thinking of a gate as being analogous to a bare
3573        function and a core as being analogous to an object or struct
3574        with methods or something with a `self` pointer
3575 
3576~rabsef Yah I need to see more non door cores to be able to understand.
3577        I lack sufficient other programming knowledge to understand
3578        fully rn I think
3579 
3580~somlus In any case thanks for the sesh and for the patience with
3581        explaining. I'll be digging in more over the next week.
3582 
3583~rabsef we should do another one next sunday maybe - I like this
3584 
3585~somlus ok I'm in for sure. Same time?
3586 
3587        I can give you more time to get home so there's no rush. 9am
3588        PST?
3589 
3590~rabsef That sounds good, yes!
3591 
3592~somlus Great! penciled that in
3593 
3594~rabsef https://pastebin.com/ac3Yc1ky
3595 
3596        let's walk through the assignment to comment this code.
3597 
3598        https://urbit.org/docs/tutorials/hoon/recursion/
3599 
3600        This gives it away but we'll frequently need to know how to use
3601        recursion in our programs so you might look @ it too
3602 
3603~somlus yes perfect.
3604 
3605~hanryc man, i love this channel
3606 
3607~rabsef I'm so stumped on Week 6 of hoon school
3608 
3609        so we may have a stumbling block lol - I will surmount
3610 
3611~hanryc Well, maybe I can help sometime this week. Which assignment is
3612        it? Ahh still generating hands of cards?
3613 
3614~rabsef yes
3615 
3616~hanryc Cool, I can probably help, although I am not sure if I can
3617        explain things as well as you do. ;) I'll hit you up tomorrow as
3618        I am off to bed now. Goodnight!
3619 
3620~rabsef The errors I get are inscrutable to me and that gets frustrating
3621 
3622        I feel like I'm close to at least generating a deck of cards and
3623        then I have an idea using snag and oust to generate a hand but I
3624        can't get there w/ the deck
3625 
3626        and I absolutely do not want to type out all 52 cards in a list
3627        - that's obviously against the spirit.
3628 
3629        that would be awesome ~hanryc, thank you!
3630 
3631        got a little closer - I can make a deck of cards
3632 
3633        https://pastebin.com/MP7r5uQm
3634 
3635        My issue was not understanding that generators have to be gates
3636        (or doors?)
3637 
3638        the correct answer is gate. the concept of a door still is
3639        elusive to me, but I get cores and gates now and how a core can
3640        have arms that are gates
3641 
3642        and importantly that a generator produces one output from all
3643        its payload, and that is the battery action of one arm, called
3644        buc
3645 
3646~hanryc To be honest, I don't think I quite understand doors yet ;) Your
3647        solution is looking good so far
3648 
3649~rabsef So I think the only differentiation I have for a door so far is
3650        that they're opened w/ |_ (barcab) and have a sample of their
3651        own there that can either be a bunt value or have a method using
3652        buccab e.g. b=_7 or whatever where that would be an @ud of 7
3653        (inferred).
3654 
3655        like I'm not saying that's the right distinction but I do think
3656        that's a distinction lol. Also you slam a door and kick a gate
3657        but I have no clue what that means lol
3658 
3659~hanryc Someone described what slam/kick difference was, but i think
3660        it's just how you use them
3661 
3662~rabsef Yah - but palfun was trying to tell me why my gen wasn't working
3663        because it wasn't a gate that is kicked the same way and I was
3664        like "yes."
3665 
3666        lol
3667 
3668~hanryc Anyway, thanks to you i re-read the doors docs (which have been
3669        updated) and I think the biggest difference is that doors have a
3670        sample (state) that is accessible to all the arms
3671 
3672        If I recall correctly, I couldn't get it working with hoon 101
3673        week 6 so I just used a core instead.
3674 
3675~rabsef I follow you on door definition and that's what I was trying to
3676        express in my limited way. Also, I am going to get this damned
3677        gen to work and then we can tear it apart here eventually
3678 
3679~hanryc Looking good so far. I had some weird issues trying to get the
3680        nested trap `|-` working the first time so it took me way longer
3681        than it should have.
3682 
3683        Looking good so far. I had some weird issues trying to get the
3684        nested trap `|-` working the first time so it took me way longer
3685        than it should have.
3686 
3687~rapfyr is there a reason ya'll don't wanna ask these kinds of things in
3688        the hoon-school chat?
3689 
3690~hanryc i either a) didn't know hoon-school existed or b) couldn't get
3691        in when i tried to join
3692 
3693~rapfyr i mean presumably some people aren't enrolled. just trying to
3694        figure out if we're doing something wrong.
3695 
3696~rabsef This room is really for trying to make the information even
3697        more basic that hoon-school allows for; and also to allow me to
3698        solidify some of my thinking by helping pass the knowledge to
3699        others. I hope that's not frustrating to your purpose
3700 
3701        I don't think you're doing anything wrong - I think that hoon
3702        is very complex for people with 0-very little programming
3703        knowledge, and this is a room where we wont annoy other people
3704        by reviewing and reviewing and remediating and reviewing
3705 
3706        Actually - I think the lessons could have more readings
3707        associated w/ them and spoon feed the user more but I suspect
3708        that's something I'd like for me, and others wouldn't like, so
3709        it's hard for me to say if that's better or just my preference.
3710 
3711        For instance - I think I need a random number generator to
3712        use mod my list length for the card selector, and while your
3713        areology video covered it, I think I need more review - reading
3714        the magic 8 ball text has helped. Unfortunately, one other place
3715        that's used has cards in it so I had to ignore that page because
3716        I don't want to cheat at all - I want to actually learn rather
3717        than just copy
3718 
3719        I can draw x number of random cards
3720 
3721        now to figure out how to make y sets of x number of random cards
3722 
3723~minder I think if all the assignments & instructions were about playing
3724        cards, I'd understand it. X^D
3725 
3726~rabsef lol ~minder :)
3727 
3728~somlus hi rapfyr - the main reasons I created hoontards are that I
3729        felt my questions were so incredibly basic that my questions
3730        were derailing other more sophisticated conversations by being
3731        inserted into them. People would respond to my simple questions
3732        with a simple explanation that might have spawned another
3733        question or was not even simple enough for me. So this is a much
3734        more dialogue-intense chat with lots of back and for and "wait,
3735        but why?" questions. Perhaps it is simply impostor syndrome but
3736        I did not feel like I belonged in the hoon-school or urbit help
3737        chats and wanted a free space to ask truly basic questions.
3738        Also, though we are using hoon school exercises, the idea was to
3739        do activities that were outside of that scope and perhaps simple
3740        and similar to hoon school exercises, but not at all the same
3741        exercise.
3742 
3743~rabsef i fugging did it. lol I made a card generator and I mostly know
3744        what's going on
3745 
3746~somlus ????    
3747 
3748~hanryc nice work!
3749 
3750~rabsef https://pastebin.com/v13RdXsR
3751 
3752        look @ this ugly bastard
3753 
3754        thanks hanryc. I’d like it if we could come up with some other
3755        challenges to keep moving forward with this. I’m at a loss
3756        tonight but I’m going to think of some and start again tomorrow
3757 
3758        also thanks somlus :)
3759 
3760~rapfyr ayyyyye great work!
3761 
3762~rabsef ok I have a challenge for myself. I want to make an ask gen w/ a
3763        loop to play war w/ 1 deck against the computer, w/ x number of
3764        cards
3765 
3766        since war doesn't really have any strategy, it should just play
3767        itself based on the cards you play
3768 
3769        That's not quite right - that wouldn't be an ask - there's
3770        something there tho. I want to be able to simply allow the user
3771        to select 1 of n cards from their hand and play them against the
3772        computer
3773 
3774        and the computer will try to do the same back, w/ higher value
3775 
3776        eventually I'd like to make a game
3777 
3778~somlus That's a great basic goal. Did you see the landscape tile apps
3779        that came out of some people's hoon 202? Not that you're that
3780        far along but they are cool
3781 
3782~risruc ~rabsef, have you read cryptonomicon?
3783 
3784        you could build a solitaire cipher
3785 
3786        https://en.wikipedia.org/wiki/Solitaire_(cipher)
3787 
3788~rabsef I have read the first part of cryptonomicon several times, but
3789        didn't finish
3790 
3791        I'll look @ that
3792 
3793        good call
3794 
3795~risruc you should read it again. it's a blast.
3796 
3797~rabsef I loved Anathem - read that?
3798 
3799~minder Oh man, that was my favorite sci-fi book until I read the Three
3800        Body Problem.
3801 
3802~risruc have not yet, ~rabsef
3803 
3804~rapfyr would having a hoon 0 week have been helpful for those of you
3805        who are in 101?
3806 
3807~rabsef What would that look like?
3808 
3809        For me, having more weeks generally would have been helpful. The
3810        jump from week 1 to even week 2 is quite a bit of intellectual
3811        ground to cover. In my experience, also, week 5->6 was a big
3812        jump. Supposing there were maybe 12 weeks the material could
3813        be further spaced out. Also, the documentation used as class
3814        material is a little over my head at times. I think a less
3815        skilled programmer could rewrite it w/ a view towards people w/
3816        less experience. At the same time I recognize that there are
3817        a ton of skilled programmers who need the documentation to be
3818        precise and deep so they can get started faster.
3819 
3820        TBH I just want to keep doing it - I want to do 201 or 202 or
3821        whatever and then more thereafter. I want to do this with you
3822        people.
3823 
3824~rapfyr mostly probably a crash course into more basic programming ideas
3825        to help people with little to no experience.
3826 
3827        thanks for the feedback on those two spots
3828 
3829~rabsef I think there would be a lot of value in the crash course
3830        element. Reviewing basic existing generators and walking through
3831        each part (like is done w/ some of the walkthroughs), first at
3832        a high level (what they do) and then down to how they do it.
3833        It's sometimes confusing that the demonstrative pieces in the
3834        texts reference doing things in dojo but the same code in a gen
3835        doesn't work - you must figure out how to do it in-gen.
3836 
3837        I would be really interested in doing the Week 0 class, still,
3838        actually. I think that I'm learning a lot but I still lack some
3839        of the fundamental ideas of programming that underlie the things
3840        that I'm doing - like I don't really understand what a state
3841        machine is; while the hoon syntax page is really good, there
3842        are some things I didn't get, like (list @) is actually a type
3843        generator, while +$ type ?(@) is also a type generator and you
3844        needn't do both (e.g. +$ type ?(list @)) in one line if you just
3845        mean list. It's hard to explain where the lack of knowledge is
3846        tho - and I'm sure its different for everyone.
3847 
3848        One more thought - I would really like to have some
3849        introductions into the existing vanes of urbit so that I could
3850        start making things that interact w/ them (for better or worse).
3851        I think I'm going to have to do a lot of learning about the
3852        language before I can read hoon effectively so as to read the
3853        vanes and make something
3854 
3855        That solitaire cipher looks cool and someone has done the math
3856        to make it more secure so obviously I have to implement that
3857        version
3858 
3859        gonna finish work and then read that and start
3860 
3861        thank you.
3862 
3863        I just practiced that cipher like 3 times to make sure I
3864        understood it
3865 
3866        I don't think this will be terribly hard - and it will help me
3867        learn map functions better - I have trouble w/ the map functions
3868 
3869        https://pastebin.com/psWsCtci
3870 
3871        here's that trouble I was talkin about - I do not understand how
3872        to call and draw from a map at all
3873 
3874        I can copy the way it's done elsewhere but (lesson 5) but I
3875        don't get why this doesn't work. I also tried declaring =/
3876        test=(map @t @ud) alphamap to no avail
3877 
3878~rovpec test
3879 
3880~rabsef rcvd somlus noon
3881 
3882        moon*
3883 
3884~somlus grazie!
3885 
3886        Rabsef, I copied the initial hoon walkthrough text from the
3887        website and commented on it. I'd like to cover the yellow
3888        highlighted portions of the comments with you. The other
3889        comments are just suggestions I will eventually submit. Can we
3890        do this on Sunday too? Or casually here?
3891 
3892        https://docs.google.com/document/d/1mqkbPVYLPjBwsYPUkdGeDDX2eQgz
3893        yZvy-H-zUpLwUPY/edit?usp=sharing
3894 
3895        (or you can actually comment in the document with that link)
3896 
3897        I went through and did some commenting - I think you bring a lot
3898        of great questions to the table and I'm excited to speak w/ you
3899        Sunday.
3900 
3901        I have a phone call I have to make at 2PST on Sunday but I can’t
3902        imagine well go that long
3903 
3904~somlus oh cool! Should we just begin as planned at 9am and cover this
3905        and then go through some code together? I'll check out your
3906        comment responses. That was fast!
3907 
3908~rabsef Yah let's do this first then we can comment the other code I was
3909        talkin about above and then we'll see what happens
3910 
3911        oh and ~palfun just answered re: =/ p q r. r is everything to
3912        do w/ the subject from there, where =/ creates a face for an
3913        element q of the subject that is then and thereafter pinned as
3914        that face for r the program
3915 
3916        that dude is a sincerely nice dude - he obliges my incessant
3917        questioning
3918 
3919~somlus ok we can look at =/ in the list example and go from there. I am
3920        still behind on vocab such as 'face' but will work on it.
3921 
3922        glad to hear he's so helpful! and thank you again for getting to
3923        that so fast. that's kind of amazing.
3924 
3925~rabsef I like engagement and this is engaging. I feel helpful and it’s
3926        helpful to me. I really enjoy this space. second to /weird maybe
3927        :)
3928 
3929~somlus /weird is pretty damn high value, that's for sure. What a gem!
3930 
3931~ronreg where is /weird hosted?
3932 
3933        ~rabsef-bicrym: based on some conversations in u-h, I think the
3934        developers aren't satisfied with the documentation for how types
3935        work
3936 
3937        i.e. it's genuinely unclear
3938 
3939~rabsef ~litmus-ritten/weird
3940 
3941        I will always be supportive of more documentation - I feel like
3942        I’m getting there but through trial and error sometimes.
3943 
3944        ~ronreg-ribdev, do you understand how to use a door to make a
3945        "state machine" as described in the bank account example
3946 
3947        Basically I'm to the point where I need to code transformations
3948        against a deck to allow for the solitaire cipher
3949 
3950        but I don't really understand how to make an arm that maintains
3951        the state of "deck" given an initial starting condition of like
3952        a sorted deck (eventually it will be either a default deck or
3953        one entered by the user as an argument
3954 
3955        https://pastebin.com/PBsysWTF
3956 
3957        ^you can see the rest of what I've done so far here if u want
3958 
3959~somlus Hey ronreg, top of the 'mornin to you.
3960 
3961        ~rabsef, always a solid presence, good morning as well. I envy
3962        you the mountain morning air.
3963 
3964~ronreg ~rabsef-bicrym: not exactly, I've been looking at that very
3965        example when I've had time over the past few days though
3966 
3967~rabsef re: stateful gen - I may be using the wrong term. What I want
3968        to do is, based on certain conditions, take a starting list,
3969        modify it, and return it to the gen, then depending on further
3970        conditions either be done or repeat the modification
3971 
3972        re: Solitaire Cipher
3973 
3974~ronreg that model of repeatedly calling a function on your state that
3975        returns a possibly-modified version of your state is pretty
3976        common in functional languages
3977 
3978        btw were people aware of https://github.com/yosoyubik/hoonschool
3979        ?
3980 
3981        ah yeah, the 1.8 bank account example is using =~, which
3982        composes an arbitrarily long list of expressions
3983 
3984        but I"m not sure what that means
3985 
3986        something about modifying the subject each time?
3987 
3988~rabsef I need to think more about how to do it. I have an arm that
3989        gives me a starting state, then I want to have an arm with
3990        transformations that I call based on whether or not something
3991        else has happened, but the part I have trouble with is how to
3992        store the modified state as the new starting state for the
3993        modifier
3994 
3995        Yah like, I could use that if I understood how to start w/ the
3996        balance after the transformations on the next go-round
3997 
3998        I tried a few times to no avail and now I'm just working on a
3999        blogpost
4000 
4001        Yah - I mean it takes the arm newaccount, gives it from the bunt
4002        value of balance 100, 100, -50, then produce resulting balance
4003 
4004        but I want to then come back to that and go "ok so I want to
4005        next time call this as balance = 150, do again.
4006 
4007        i.e. 300 output
4008 
4009        further, I want to allow for an optional card list given at the
4010        beginning of the gen to set your own starting state, but I'd be
4011        ok w/ the default state only if I could figure this other part
4012        out
4013 
4014~ronreg I'm looking at the 1.8 code now trying to figure this out
4015 
4016        I believe there's a way to set a non-bunt value as the initial
4017        value of the door sample
4018 
4019        lets see, the whole thing is in a =<
4020 
4021        that's backwards =>
4022 
4023        [%tsgr p=hoon q=hoon]: compose two expressions. Produces the
4024        product of q, with the product of p taken as the subject.
4025 
4026~rabsef I'd thought about using a counter outside to arbitrarily count
4027        the first cycle and use that as an argument going into my other
4028        door/arm but that seems really lazy/wrong
4029 
4030        and, by that argument, ?: to either a first transformation or a
4031        second one, and so on.
4032 
4033        where first transformation reads from default and subsequent
4034        read from an "output state"
4035 
4036~ronreg it makes sense to me that you should be able to do something
4037        analogous to the bank account door example, only setting the
4038        initial value from a user-supplied input, and then repeatedly
4039        updating it
4040 
4041        without any kind of counter magic
4042 
4043        https://urbit.org/docs/tutorials/hoon/bank-account/
4044 
4045        "We've chosen here to wrap our door in its own core to emulate
4046        the style of programming that is used when creating libraries."
4047 
4048        I don't think tha'ts what you're doing in your linked code
4049 
4050~rabsef I'm not, yet at least.
4051 
4052~ronreg so that might be the trick, have one of the arms of your outer
4053        core return a door that has some additional arms on it
4054 
4055        I'm trying to figure out exactly how the bank acct works
4056 
4057~rabsef Let's say I have starting list ~[1 2], I want to start there
4058        then maybe flop it and add 1 to each. Then the next time I call
4059        that arm, I'm confused as to how I'd say - don't use starting
4060        list
4061 
4062        without something counting how many times it has operate before
4063 
4064~ronreg so your starting list ~[1 2] is analogous to the starting
4065        balance in the door that new-account returns
4066 
4067~rabsef yep
4068 
4069        oh and importantly, once I do a transformation, I need to
4070        operate on a nother part of the subject, before I do the
4071        transformation again
4072 
4073~ronreg and then, in the docs example, =~ makes it so all the subsequent
4074        invocations of deposit and withdraw are changing the state of
4075        that starting balance
4076 
4077~rabsef so we have to come out of that arm, then go back in
4078 
4079~ronreg I'm still not clear how that is guaranteed
4080 
4081~rabsef maybe that's where I'm wrong - we don't need to come back out
4082        if I do all the operations there and output form that arm,
4083        including the transformations to other parts of the subject
4084 
4085~ronreg maybe you need a more complicated sample, like with multiple
4086        copies of the list, and you're only working on one at a time?
4087 
4088~rabsef that's interesting too
4089 
4090        possibly lol :)
4091 
4092~ronreg this syntax seems important: +>.$
4093 
4094        "+> is wing syntax. This particularly wing construction looks
4095        for the the tail of the tail (the third element) in $, the
4096        subject of the gate we are in, which is the entire new-account
4097        door. We change balance to be the result of adding balance and
4098        amount and produce the door as the result. withdraw functions
4099        the same way only doing subtraction instead of addition."
4100 
4101        lol typo in the docs :)
4102 
4103~rabsef yah so I thought maybe I could like set +>.armname but I tried
4104        it and it didn't like it
4105 
4106        I may have fucked it up tho
4107 
4108        I could try again
4109 
4110~ronreg ah yeah so the syntax here is $(variable-name new-value)
4111 
4112        so each call to withdraw/deposit is calling $ but with balance
4113        (the name of the sample) updated to something else
4114 
4115        except it's not $ it's calling but the +> child of $
4116 
4117        still not sure what that does
4118 
4119~rabsef Somehow, $ refers to newaccount and +>.$ is balance
4120 
4121        I can't for the life of me figure out how that's determined. It
4122        doesn't look like the tail of the tail to me
4123 
4124        It looks like the head of the tail, if the head is |_
4125 
4126        remember the wing notation is - or < alternating for head and +
4127        or > alternating for tail
4128 
4129~ronreg ah okay I think that makes sense
4130 
4131~rabsef Do you understand how balance is tail of tail? I can't get it
4132 
4133~ronreg new-account is an arm of the outer cell, so anywhere within it $
4134        should be valid
4135 
4136        $ is the "self" of new-account, not the door... i think
4137 
4138        it would be good to check the exact definition of a core arm
4139 
4140        https://urbit.org/docs/tutorials/hoon/arms-and-cores/ might be
4141        helpful
4142 
4143        Cores, Gates, and Traps
4144 
4145        so maybe it's the other way around - $ is a special name created
4146        by the door, not the core. and the core is the +> child of the
4147        door's subject?
4148 
4149~rabsef tbh that section and 1.8 were hard for me to understand
4150 
4151        like the part about modifying a core and having to save it back
4152        to itself was pretty hard to understand imo
4153 
4154~ronreg yeah I don't quite grok it yet either
4155 
4156        I think that it's not so much "saving it back to itself", as
4157        creating a new subject where the same name points to a modified
4158        copy of itself
4159 
4160        that's a common functional programming pattern
4161 
4162        there's like an implicit stack of states
4163 
4164~rabsef that makes sense - that's what they're trying to tell us in
4165        recursion
4166 
4167        ~palfun says ~palfun-foslup. you'll want your arms to produce
4168        the updated thing, ~14.47.01
4169 
4170        probably. (or the updated core, if there's many things
4171 
4172        you want to change). look at code that uses =^
4173 
4174~ronreg I need to go grab some food but afterwards I'm gonna look at
4175        this again
4176 
4177~rabsef yah my laptop died and I’m going to hit some chores. talk to you
4178        soon. thanks for thinking thru this with me
4179 
4180~hanryc In that example `new-account` is the door
4181 
4182        And the $ refers to the gate of the say generator `|= *` which
4183        has a head `*` and a tail which starts with the `:-` rune
4184 
4185        .
4186 
4187        The `:-` rune is a cell that has a head `%noun` and a tail which
4188        is the rest of the code (produces the door). So, the tail of the
4189        tail of the generator $ is the door itself, `new-account`
4190 
4191        the expression `+>.$` is `new-account` door and so the arm
4192        is returning the door with a modified subject `(balance (sub
4193        balance amount))`
4194 
4195        Similar to "Modifying a core's payload" section on the arms and
4196        cores page. There's an example `c(a 99)`
4197 
4198        Thanks for talking this through, btw. I didn't understand this
4199        the first time, but I think I understand it much better now that
4200        I read your conversation. lol
4201 
4202~somlus First we will go through some of the annotations & discussion
4203        on:
4204 
4205        https://docs.google.com/document/d/1mqkbPVYLPjBwsYPUkdGeDDX2eQgz
4206        yZvy-H-zUpLwUPY/edit?usp=sharing
4207 
4208        and then we'll do whatever simple code that rabsef created. FYI
4209        ~rabsef I no longer have scrollback, so you will have to re-post
4210        the paste.bin for us to go by...
4211 
4212~rocted .
4213 
4214~rovpec .
4215 
4216        (prepping my moons, please disregard)
4217 
4218        Ok rabsef, no rush. LMK when you're up and running.
4219 
4220~rabsef I'm here - your planet just doesn't like it when I talk - this
4221        will show up momentarily
4222 
4223        I arrive!
4224 
4225~rovpec haha!
4226 
4227        welcome
4228 
4229        Moons and also my planet can be quirky
4230 
4231        I got a laptop with a full Ubuntu install now, and it screams.
4232        Somlus doesn't live on it but moons do.
4233 
4234        Do you want to look at the document Q&A first?
4235 
4236~rabsef It happens. I saw you posted some surreplies to my replies.
4237        Should we just start at the top?
4238 
4239~rovpec perfect.
4240 
4241~rabsef just to remember me, you've successfully made generators before
4242        right?
4243 
4244~rovpec yes, well you and I made the list together and also added some
4245        numbers. pretty basic stuff.
4246 
4247        Oh and also as a reminder, we should try to focus as much as we
4248        can on each concept, without (for now) explaining something with
4249        lots of hoon terms.
4250 
4251~rabsef Ok so let's start at that top line w/ bartis. We have an
4252        explanation of a later question right in the definition of
4253        bartis in the reference
4254 
4255~rovpec cool
4256 
4257        |=
4258 
4259~rabsef That is |=
4260 
4261        https://urbit.org/docs/reference/hoon-expressions/rune/bar/
4262 
4263        "The product of |= is [a hoon function] where p defines the
4264        gate['s input] and q is [the rest of the gate resulting in
4265        output]
4266 
4267~rovpec Aha now I see where $ is mentioned.
4268 
4269        It is the work product, of the bar cores
4270 
4271~rabsef yep - "A gate is a core w/ one arm named $. I'm not sure I know
4272        exactly what a core is outside of saying a gate is a form of a
4273        core, a door is a form of a core, and a core is something like a
4274        collection of things to do
4275 
4276~rovpec ok that's helpful actually.
4277 
4278~rabsef And to clarify p and q, the spec for |= is |= p q
4279 
4280        also you're right about the double space - that gets me a lot in
4281        my coding
4282 
4283~rovpec Let's dig into that 'the product of' line a bit.
4284 
4285        oh! was just about to ask that
4286 
4287        do you have a link to that spec?
4288 
4289~rabsef nearly everything in "tall form" meaning not using irregular
4290        form or long form in ()s has 2 spaces between operant pieces
4291 
4292        |= a=spec b=hoon
4293 
4294        it's in the hoon-expressions link above - did that come thru?
4295 
4296        yah I see it
4297 
4298~rovpec spec means "gate's input"?
4299 
4300        yes it did, stand by
4301 
4302~rabsef That's a good question. I think it actually means specification
4303        of argument - some things that are not 'input' can still be
4304        specifications
4305 
4306        an argument can some from somewhere else or be declared if that
4307        makes sense
4308 
4309~rovpec sorry had to scroll down a lot, got to the bartis section.
4310 
4311        yes it makes sense
4312 
4313        looking at bartis now
4314 
4315~rabsef if you look @ the examples, check out how one is in long form
4316        and has single spacing, while the tall form has double
4317 
4318~rovpec what is "AST"?
4319 
4320~rabsef long means like it sounds - across rather than down
4321 
4322~rovpec yes will do, ok just a sec.
4323 
4324~rabsef "Abstract Syntax Tree"
4325 
4326        I'm not sure how that is intended to be different from the
4327        regular, frankly
4328 
4329~rovpec ok will disregard that.
4330 
4331        looking at long versus tall form
4332 
4333        ...
4334 
4335~rabsef I unfortunately lack some of the required cs knowledge for
4336        interpretation
4337 
4338~rovpec no worries. I do not see a discrete example there of a long /
4339        tall form comparison, so let's move on and I will look closer at
4340        long and tall form equivalencies later
4341 
4342        Thanks for pointing me there. I now see that $ is just the end
4343        of |= as defined by the |= function
4344 
4345        I was wondering about $.
4346 
4347~rabsef we can do that quickly |=(a=@ +(a)) means give me an atom and
4348        increment it by one
4349 
4350~rovpec ok
4351 
4352~rabsef so tall form it would be |= a=@ <carriage return> +(a)
4353 
4354        the first being long form
4355 
4356~rovpec ok. I expected two spaces between @ and +
4357 
4358~rabsef Yes - in tall you have effectively 2 spaces insofar as you have
4359        a carriage return (confusing) but in long, within the ()s only
4360        one
4361 
4362        anyway, let's talk about children - |= takes two, can you
4363        identify what they are?
4364 
4365~rovpec ok, got it. long, one space only. tall, two spaces or <CR>
4366 
4367~rabsef precisely
4368 
4369~rovpec yes I now finally can. the first chile is defining the end=@ (or
4370        naming variable "end" with the user input, interpreted as a real
4371        number)
4372 
4373~rabsef (incidentally, somewhere in the documentation they say that hoon
4374        only cares about two types of spaces - either 1 space or more
4375        than 1 space)
4376 
4377~rovpec and the other "child" is actually multiple lines of code up to
4378        and including the line that begins with $
4379 
4380        (got it on the space comment)
4381 
4382~rabsef (so I think you could do 10 or a cr or whatever, but I haven't
4383        tried it)
4384 
4385        right - now that will be relevant when we get to :- in the
4386        formula we're looking at
4387 
4388~rovpec ok
4389 
4390~rabsef let's make a note about "faces" - a face is just a name for a
4391        part of hoon, it can resolve to data (leg) or more code (arm)
4392 
4393~rovpec a defined part, or an arbitrary part that is used for some
4394        discussion of an overall code?
4395 
4396~rabsef end is the face of our input, or the name of it, which is, in
4397        this case, data
4398 
4399~rovpec "part of hoon" seems ambiguous since it could be part of a
4400        program, or part of the overall hoon language
4401 
4402~rabsef a defined part - ti's really a part of the subject. But all hoon
4403        is part of the subject as is all data.
4404 
4405~rovpec hmm, not following that
4406 
4407~rabsef ok - it's not incredibly important. The important part is that a
4408        name (face) can mean a static data point, or more code to run
4409 
4410~rovpec ok
4411 
4412        perfect, thanks - plain english is super helpful for my stage
4413 
4414~rabsef ok so we're basically thru line one. Let's look @ 2. 2 children
4415        here too =/ specification value
4416 
4417        And this is where it gets more tricky but also very important
4418 
4419~rovpec barfas
4420 
4421        ok
4422 
4423        tisfas
4424 
4425~rabsef Since hoon expressions have determinate numbers of children - we
4426        can read where an expression ends by knowing how many children
4427        it takes
4428 
4429        whereas in excel, we must close things with )
4430 
4431~rovpec yes, as long as we fully understand what defines a child.
4432 
4433        but at the beginning, that is not crystal-clear
4434 
4435~rabsef agree - but now that we've established some of that we can start
4436        trying to read teh code by its expected children
4437 
4438~rovpec But I am definitely following you
4439 
4440        yes agreed.
4441 
4442~rabsef Ok - so this next line, line 3, this is where I got into trouble
4443        early on. In a lot of code I've dealt with, you can have a
4444        binary distinction - if x do y if not x, end.
4445 
4446        but hoon doesn't just let a decision branch to /end/. You have
4447        to always produce / do something
4448 
4449~rovpec yes, that's what I am comfortable with also
4450 
4451~rabsef The |- takes one argument, being again everything beneath it,
4452        and (we know now) must result in some production of a value or
4453        action to take
4454 
4455        the action to take, in turn, must always result in eventually
4456        some product
4457 
4458~rovpec got it.
4459 
4460        Can you elucidate "you have to always produce / to do
4461        something"?
4462 
4463        never mind
4464 
4465~rabsef We know here that our example will result in the production of a
4466        count of numbers, but it becomes more complex later
4467 
4468~rovpec I thought that you meant you literally have to produce forward
4469        slash
4470 
4471        sorry
4472 
4473        yes, still following
4474 
4475~rabsef ah gotcha - but let me give an example quickly. In some part
4476        of code I made, I had a ?: (branch on %.y or %.n [true false]
4477        of some other code to a or b action). I was using that inside
4478        something that needs a terminator (some runes have indeterminate
4479        children, and you have to end w/ ==) I tried to make it ?:
4480        condition == other code
4481 
4482        where if the condition were true, the program would just
4483        terminate
4484 
4485        that is failure to produce. == is not an action/product
4486 
4487~rovpec got it, thanks for explaining that. so in that case how did you
4488        work in a termination?
4489 
4490~rabsef I rewrote the formula to not have that there - I don't remember
4491        exactly - I think I put the indeterminate member inside of the
4492        ?: but I don't recall exactly
4493 
4494~rovpec no worries, but now I understand that the children must do or
4495        produce something
4496 
4497~rabsef ok so |- just takes code that will recurse back to itself - the
4498        instant problem we should see and understand is that if we don't
4499        have a defined end to that recursion we will make an infinite
4500        loop
4501 
4502~rovpec correct.
4503 
4504~rabsef Also, I think technically =/ and |- are the only parts of q of
4505        |= if that makes sense, since everything below |- is part of |-
4506 
4507        which might be helpful to think of so you don't overlap what
4508        children are whose
4509 
4510~rovpec actually yes that's really helpful.
4511 
4512~rabsef I think most technically, the hproduct of the completion of |-,
4513        considering =/ and the input value are the q term
4514 
4515        but that's 6 of 1
4516 
4517~rovpec I'm even adding more notes to the google doc, to see if this
4518        helps them do this right in the next edition
4519 
4520~rabsef Ok, so we have a declared input, a declared constant, and a
4521        process that should, taking into consideration those things,
4522        result in a value for |= so it may terminate
4523 
4524~rovpec yes, a rose by another name so they say.
4525 
4526        ok
4527 
4528        yes.
4529 
4530~rabsef So - one problem w/ hoon, or maybe not a problem, but a reality,
4531        is that its not always sure what it's producing if you don't
4532        tell it
4533 
4534        basically, since as you rightly state anything is just an
4535        increasingly long string of numbers, sometimes the types are
4536        non-obvious to the compiler
4537 
4538~rovpec really? that seems odd
4539 
4540~rabsef this results in problems if you don't tell it what you're doing
4541 
4542        I have faith that they know why they do that :) lol
4543 
4544~rovpec yeah! some of this you just tell youself "yeah but it's a
4545        deterministic computer that's why"!
4546 
4547~rabsef But this brings us to line 4. ^- which declares a type for that
4548        below it
4549 
4550~rovpec ok thanks for spelling that out. Honestly I didn't know why they
4551        needed to do that
4552 
4553~rabsef here is something I'm still struggling with ^- does not have to
4554        be a type for a product. You can declare a type of function with
4555        it too, if your hoon below that results in some sort of function
4556        to run.
4557 
4558        I don't really understand what that means but just know that ^-
4559        is not only for "values"
4560 
4561        early onhowever, we will use it for values
4562 
4563~rovpec ok thank you. super helpful for me to read what you also
4564        understand and what you do not
4565 
4566        (incidentally I have to tell you that this particular session is
4567        sort of breakthrough-y for me)
4568 
4569~rabsef ^- in our case here gives us a form for our output of |- which
4570        should be, pursuant to our goals, a list of numbers counting up
4571        and terminating in our input value
4572 
4573        fuck yah dude. that's great
4574 
4575        a list of numbers here is just a list of atoms, which is
4576        theoretically not the greatest 'cast' (identification of type)
4577        because I mean 'thisisanatom' is as much an atom as 3 or 0b1001
4578 
4579        or whatever.
4580 
4581~rovpec (expounding a bit more, the issue for me is "rules" and "flow" -
4582        the documentation presupposes that you already understand those
4583        rules and flow and does not explain them in an intuitive way, so
4584        it just remains so confusing)
4585 
4586        yes
4587 
4588~rabsef I agree - I think I have slightly more background that allows me
4589        to sort of intuit some of the rules and flow, but still far less
4590        than I need to really get the docs
4591 
4592~rovpec but is simple and they had not talked about auras yet
4593 
4594~rabsef yes right - we might change our argument there to @ud if we were
4595        imporving this
4596 
4597~rovpec got it
4598 
4599~rabsef ok now the fun part
4600 
4601        and a better discussion of children
4602 
4603~rovpec ok
4604 
4605~rabsef ?: is a branching function that has 3 children. A condition, a
4606        result if yes/true, a result if no/false
4607 
4608        what are our children respectively? - let's start w/ the
4609        condition
4610 
4611        The condition is fairly intuitive to read here =(end count)
4612 
4613        sometimes it can be harder like [assume gth means greater than,
4614        which it does] (gth 6 7)
4615 
4616        I find that hard to read, as the comparison comes first
4617 
4618        I read it now as 6 is gth 7, which is false
4619 
4620~rovpec cool
4621 
4622~rabsef we could read =(end count) as end is equal to count, which
4623        is potentially different from =(count end) though not really
4624        potentially because of the transitive property or whatever
4625 
4626~rovpec (missed that but no worries)
4627 
4628~rabsef also, and incidentally, =() is a long form of ?= and also an
4629        irregular
4630 
4631~rovpec good to know thanks
4632 
4633~rabsef for our purposes =(end count) will give us a %.y or %.n based on
4634        whether it is true
4635 
4636        then ?: will check that output and say ok, if %.y then my second
4637        child, if %.n then my third
4638 
4639        let's skip ~ for a second
4640 
4641~rovpec correct, though you never actually visually see the %.y or %.n
4642        here, you just have to know that logic behind the scenes is the
4643        2nd child on y and 3rd child on n part
4644 
4645        ok let's skip it
4646 
4647~rabsef The third child, the no condition, is yet another rune.
4648 
4649        (to your point there if we just had ?= we would get, in output,
4650        our %.y or %.n assuming we commented out all the rest of the
4651        code
4652 
4653~hanryc Did you two see my discussion of the door example from
4654        yesterday? I put it up after you left
4655 
4656~rabsef I did ~hanryc and I still need to think about it more
4657 
4658        but thank you for doing that
4659 
4660~rovpec :- is the "construct a cell" rune
4661 
4662~rabsef Right - and what kind of cell
4663 
4664~hanryc ok, i was just making sure it didn't scroll off
4665 
4666~rovpec (thank you for the clarification on y and n)
4667 
4668        what kind of cell....
4669 
4670~rabsef [%clhp p=hoon q=hoon]: construct a cell (2-tuple).
4671 
4672~rovpec a list of real number atoms
4673 
4674        since there was (list and also @)
4675 
4676~rabsef explicitly a 2 part cell. This is why it's important. :- has
4677        determinate children
4678 
4679        2 eto be exact
4680 
4681~rovpec oh, ok, yes that too.
4682 
4683~rabsef the first child is easy to read, it's count
4684 
4685        the second child is less easy to read - and this is where $
4686        comes in. The second child is the entirety of every subsequent
4687        potential recursion until we break the loop
4688 
4689        Let me lay this out a little visually if I can - let's pretend o
4690        is the name of our output each recursion after the first, so o1
4691        is recurion 1 o2 is recursion 2 etc
4692 
4693        we know we start with count=1
4694 
4695~rovpec ok
4696 
4697~rabsef our cell that :- creates is then, [count=1 [o1=2 [o2=3 ...
4698        [oN=end ~]]]
4699 
4700        This is because on every recursion until count = end, we form
4701        yet another twouple cell
4702 
4703~rovpec is it because a duple cell which adds 3 to [1 2] is [1 2 3]?
4704 
4705        the two things in the duple are 3 and the already existing unit
4706        [1 2]?
4707 
4708~rabsef Let's walk through it - :- count=1 creates the first half of a
4709        duple cell
4710 
4711        [1 ....
4712 
4713        the recursion says "add one to count and start from |-"
4714 
4715        so to find our second half of the duple, we have to do that - so
4716        now count is = to two and lets say our end value is 3 for the
4717        sake of argument
4718 
4719        [1 ..... (is 2=3? no, so) the second half of our first cell, is
4720        the start of a new cell
4721 
4722        [1 [2....
4723 
4724        that is we go back to line 7 for the second half of the first
4725        cell
4726 
4727        [1 [2 .... (recurse again, is 3=3, yes, so) ~]]
4728 
4729        [1 [2 ~]]
4730 
4731        aka ~[1 2] or [1 2 ~]
4732 
4733        does that make sense?
4734 
4735~rovpec Ah, so interesting it's not actually building a list, it's
4736        creating a cell inside a cell inside a cell and are are
4737        open-ended duples until they all slam closed at the same time
4738 
4739~rabsef a list is a null terminated cell, and yes, so technically "no,
4740        but yes" to your relaying of it
4741 
4742        lol
4743 
4744~rovpec lol
4745 
4746~rabsef but yes exactly, until every cell has its second half, they're
4747        all sort of floating and recurring
4748 
4749~rovpec and then you were going to tell me - how does it know that ~ is
4750        the end not only of the list but of the entire gate?
4751 
4752~rabsef because it is the deterministic duple second half w/ a value of
4753        null
4754 
4755        if end=3, our second half of [1 is intedermined until we
4756        recurse.
4757 
4758        Imagine this what if we put 1 as end?
4759 
4760        what happens?
4761 
4762        we only get ~ right?
4763 
4764        ~ is a value so we don't even go to line 7, then the ^- casts it
4765        as a list of @ so we just get [~]
4766 
4767        in any number end > 1 we actually start forming cells, that will
4768        by logic always end in a cell of [count(x) ~]
4769 
4770        for the same reason
4771 
4772        Another way of saying this is that, on the first run, everything
4773        below |- is |-'s children, but as soon as we get to $, if we do,
4774        the subsequent recursion is actually a child of :-
4775 
4776~rovpec is it safe to say that if there is a null value in a list, then
4777        the system automatically knows "this list is finished, there
4778        is nothing more to do, I have produced the final product by
4779        definition" and so it ends, and you do not need to tell it to
4780        end?
4781 
4782~rabsef it's nested
4783 
4784        no - that's not quite right. The system knows that the list is
4785        done for 2 reasons:
4786 
4787        1) :- only takes two children, the first of which will be count,
4788        but the second is indetermined until ?: kicks again
4789 
4790        if ?: says %.n, then the second half of the cell is another
4791        recursion, and we must process again
4792 
4793        and 2) if ?: says yes, then the second half of the recursion is
4794        ONLY line 6, which is the value ~
4795 
4796        we could replace line 6 with other code and not cast this as a
4797        list and just get a cell that is not null terminated (end with 0
4798        for instance) so if we put end = 3 we would get the cell [1 [2
4799        0]]
4800 
4801        does that make sense?
4802 
4803~rovpec is it by definition that another recursion cannot be a 2nd
4804        child? that a 2nd child can only be what comes after the
4805        recursion is complete? Sorry I'm having a hard time with this,
4806        even though you have spelled it out clearly
4807 
4808~rabsef Let me try saying it this way -> The children of a cell must
4809        always be computed down to a value (not sure if this is
4810        perfectly true, but assume it here)
4811 
4812        on our first go through of this program, we give one value and
4813        one piece of program that must be computed down to a value
4814        before it can be accepted as a child
4815 
4816        Thus, the second half of every cell created by execution is
4817        another recursion, until end=count. On the end=count run, the
4818        second half of the cell is just ~
4819 
4820        (which is a value)
4821 
4822~rovpec ok, I will ruminate on that. I do think that if I understand
4823        the entire flow of this particular generator, I will 1) have a
4824        framework for understanding hoon syntax and 2) be able to help
4825        them create an initial part of documentation that can take even
4826        early learners through thisphase.
4827 
4828~rabsef let's say it this other way w/ sandwiches
4829 
4830~rovpec I'mportant to know that firmly the children of a cell must
4831        always be computed down to a value.
4832 
4833~rabsef imagine I'm making a multi layer sandwich
4834 
4835~rovpec ok
4836 
4837~rabsef the first half of each layer is going to be bread. bread + what?
4838        I have to look at my instructions - lets say it's n levels tall
4839 
4840~rovpec ok
4841 
4842~rabsef until I am at n, I will have to add an ingredient other than
4843        bread, and then a bread, and then look at my instructions again
4844 
4845        bread butter bread jam bread meat bread lettuce bread mayo bread
4846        ~ <- air, my last bread is the run where the next instruction
4847        will be n layers which means nothing on top
4848 
4849        count is bread always, and our recursion is our look @
4850        ingredients / our ingredients
4851 
4852~rovpec This structure makes sense
4853 
4854~rabsef [1 (look @ ingredients and the next ingredient is) [ then bread
4855        then look at our ingredients and our next ingredient is [ then
4856        bread
4857 
4858        and so on
4859 
4860        or [1 [2 [3 ~]]]
4861 
4862        ignore that fact that we have an ingredient on bottom lol
4863 
4864~rovpec haha some sandwich!
4865 
4866~rabsef we might even better define [ as bread so :- starts w/ [(bread)
4867        then our current ingredient which is count, then [ count [ count
4868        until we're done and then ~]]
4869 
4870        but does that help at all?
4871 
4872~rovpec i was thinking on [ as bread anywasy so all good!
4873 
4874        ok thanks. i'm sorry we did not get to the other thing you
4875        wanted to cover but I think I have enough to go on here to dig
4876        into the final understanding of this generator.
4877 
4878~rabsef No this was great - do you want to call it there or would you
4879        like to keep going?
4880 
4881~rovpec I think that for me the best course of action is to review and
4882        review and be sure I get this. Can we do your plan for next
4883        Sunday? remind me of what it was and I will put that into my
4884        appointment and prepare
4885 
4886~rabsef Yah I actually can't get scrollback any more either so I'll
4887        look @ the notebook for hoon school - I think it was the second
4888        assignment to just comment some code like this that does a
4889        factorial output
4890 
4891        same deal there - we will want to walk through recursion again
4892        and see how it works in a different context
4893 
4894        https://urbit.org/docs/tutorials/hoon/recursion/
4895 
4896        let's do that
4897 
4898~rovpec ok fantastic. This speed and level of detail is amazing! I am
4899        indebted
4900 
4901        ok I will put that link in my calendar.
4902 
4903~rabsef Let's just remember that I don't know nearly enough to be
4904        considered even a moderate user, so we're getting big concepts
4905        but may have details excluded
4906 
4907        I don't want you to be misled by something I say
4908 
4909~rovpec no worries, all good. We can refine it as we go.
4910 
4911        Ok I'll do another hour or so of self study now. Thank you again
4912        rabsef!!
4913 
4914~rabsef outstanding - I'll see you sunday then. Sunday school w/ hoon!
4915 
4916        :)
4917 
4918        thanks for having me - talk to you soon
4919 
4920~rovpec thanks! talk soon
4921 
4922~dinlut sorry about the random "j" comment in the google doc. the
4923        floating action button to make a comment was just above the
4924        right nav button, must have clicked it by accident
4925 
4926        interacting with docs on a phone is not ideal
4927 
4928~somlus no worries at all dinlut. seriously feel free to add any
4929        comments or responses. I'm trying to create a version that
4930        speaks to very early non-coder learners, so I'm trying to
4931        identify all that is not clear with the beginning parts of the
4932        documentaiton.
4933 
4934        Hey ~rabsef I put our chat into a document. not sure
4935        if it has value to you but it might. Here it is:
4936        https://www.dropbox.com/s/by8b1jvpmpjsson/Rabsef%20on%20list%20g
4937        eneration%2001-12-2020.docx?dl=0
4938 
4939        (btw no small task to copy paste landscape chat into a
4940        well-ordered document! lots of workarounds needed)
4941 
4942~hastuc hey ~somlus if you want to get the json state of your current
4943        chats
4944 
4945        here's a generator
4946 
4947        save it in /gen/chat-backup
4948 
4949        https://pastebin.com/YFVZD6LH
4950 
4951        ~somlus save this as /gen/chat-backup
4952 
4953        https://pastebin.com/YFVZD6LH
4954 
4955        */gen/chat-backup.hoon i mean
4956 
4957        and then run it like this
4958 
4959        *%/chat-backup/json +chat-backup
4960 
4961~rabsef ~hastuc - At your leisure some time I'd like to talk to you
4962        about the generator I'm working on for this solitaire cipher. I
4963        am just not getting it
4964 
4965~hastuc yeah for sure i'm available for the next 8 or so hours lmk when
4966        you want to talk
4967 
4968~rabsef also that's very dope
4969 
4970        nice job
4971 
4972~somlus ok awesome ~hastuc, thank you very much
4973 
4974~hastuc just fyi it will probably hang your urbit for a couple of
4975        seconds but should be fine
4976 
4977~somlus understood, much like mounting sometimes does.
4978 
4979~rabsef ~hastuc-dibtux - I got two transformations completed yesterday
4980        - one behaves in a way that is confusing to me so I'd like
4981        to look @ it with you some time, but only after I finish the
4982        transformations. You are the real deal man
4983 
4984        (or lady - I mean it in the colloquial sense)
4985 
4986~hastuc congrats!
4987 
4988~somlus hey rabsef you need any more planets to give out?
4989 
4990~rabsef hmm I could do a small giveaway on twitter plus I do have a
4991        buddy who wants to try the raspberry pi install... if you don’t
4992        mind I’d be into it. I know you said don’t be timid but here we
4993        are lol
4994 
4995~mocrux Ah, is ARM support already implemented?
4996 
4997        If I find time I might try booting a ship on a spare rock pi I
4998        have laying around.
4999 
5000~widbur between urbit and urbit-worker, just the processes weigh in at
5001        ~750MB on my machine
5002 
5003        presumably it also consumes pagecache for the 2GB image
5004 
5005        might be a tight fit on a pi
5006 
5007        i've got an early production test version of a pine64 board from
5008        when i did alpha testing for them, if there's arm support i
5009        might try that as it's got 4 gigs iirc
5010 
5011        although finding that in my office would take an almighty grep
5012 
5013~mocrux I wonder what kind of a userbase will be required before
5014        purchase-plug-setup-and play Urbit boxes based off of these low
5015        powered arm based boards will be in demand/desired by users.
5016 
5017        I imagine you could probably include an invite and some sort of
5018        streamlined process on first boot so that they set up the ship
5019        with their own fresh keys.
5020 
5021~widbur i don't think userbase is the limiting factor, usefulness is
5022 
5023~mocrux Those are two sides of the same coin I suppose. With usefulness
5024        comes new users, which drives demand, and thereby makes it
5025        possible for such nodes to be worth developing/selling.
5026 
5027~minder The comet has a good point.
5028 
5029~widbur there are plenty of useful things you don't require a large
5030        userbase for
5031 
5032        and for others, an internet->urbit gateway would suffice
5033 
5034        for me, the question of whether something complex and useful can
5035        be built on top of urbit is not at this moment decidable with
5036        reasonable effort
5037 
5038        largely because the things i need to understand to make an
5039        educated guess are hidden behind a wall of gobbledygook jargon
5040 
5041        it takes a lot to make a functional VM-as-OS project completely
5042        unevaluatable without massive effort for a kernel dev familiar
5043        with functional programming, but urbit succeeds at that
5044 
5045        i'm not being hostile or unreasonably critical btw, the bits I
5046        do understand at an architectural overview level are really cool
5047 
5048~hodler I have to agree about the urbit jargon... I find it very
5049        difficult to learn.
5050 
5051~widbur the networking though - composability, the price being paid for
5052        idempotency-as-an-exactly-once mechanic and so on, i have 0 clue
5053 
5054        it's not a problem if this is your first programming experience,
5055        you'd have had to learn the same volume of jargon anyway
5056 
5057        but it's not as if functional programming, overlay networks,
5058        transparent messaging and so on are brand new things - these
5059        are established fields, with established terminology, and that
5060        terminology has a purpose - letting those who have already
5061        invested the time and effort to accumulate clue use it to grok
5062        new projects
5063 
5064        i've read yarvin's claims that this is all completely new, urbit
5065        doesn't owe anything to 20th century CS and so on, but it's
5066        blatantly not the case. there are some new-ish things, like jets
5067        (although an isomorphic equivalent might exist that i don
5068 
5069        't know about, since i'm not a hardcore fp guy), and the
5070        composition is new, but the building blocks very much are not
5071 
5072~ravsur The overt reason (I can't find a ref but it's out there
5073        somewhere) for the ab initio jargon is that many Urbit entities
5074        differ subtly but potentially misleadingly from existing CS
5075        concepts. The covert side, in my opinion, is about 50% an
5076        attempt to regulate the rate of adoption, so as to avoid
5077        dilution of the foundational principles, and 50% whimsy
5078 
5079        Or things indistiguishable from whimsy from the outside
5080 
5081~widbur CS has had 0 problems dealing with 'largely similar with an
5082        important difference' for quite some time now
5083 
5084        particularly in languages
5085 
5086        as a polyglot, it's something i trust my brain to handle easily
5087 
5088        whimsy can be seen in the naming, but i wouldn't say it's the
5089        reason
5090 
5091~ravsur I note that the four-letter naming scheme has slowly been
5092        dropped over the last year, so the direction of travel does
5093        appear to be away from whimsy
5094 
5095~widbur it's not a problem that it's four letters, it's that
5096        straightforward CS concepts are renamed and no write-up exists
5097        that explains networking, for example, well enough so I can
5098        figure out whether urbit performing badly even for chat is an
5099        implementation problem, design problem or both
5100 
5101        last time i brought this up ~a year ago, i was told that it was
5102        an implementation problem and a networking rewrite was in the
5103        works. clearly it's improved things massively, but not enough.
5104 
5105        is it hopeless? can i help? i can't know, apparently, because
5106        there's no usable info and RTFCing is not a practical option
5107 
5108~master ~widbur, i'm curious how you've formed the conclusion that "it's
5109        blatantly not the case"
5110 
5111~widbur since there are people in this channel who haven't programmed
5112        before learning fp, clearly it's a net positive in the world and
5113        has some excellent folks in the community
5114 
5115        ~master, because some of the bits I did manage to grok map very
5116        nicely to established concepts I'm familiar with?
5117 
5118~master from a high enough level of analysis, anything computational can
5119        be considered equivalent to anything else ... but the details
5120        matter
5121 
5122~widbur we've had some isomorphic equivalent of this conversation
5123        before, ~master, it's just that i have to create a new comet
5124        every time i check out urbit so there's no continuity
5125 
5126        in a sense, it doesn't matter what my criticisms of the
5127        terminology decisions are, the decisions have long been taken
5128        and what is, is
5129 
5130~master i believe it
5131 
5132        that may be, but there's still dramatic leeway in pedagogy
5133 
5134~widbur when i look at a cool project, and urbit is very, very cool, i
5135        want to know: 1) is it broken in a fundamental way at design
5136        stage preventing it from being useful when mature, 2) is it
5137        worth me getting involved, and 3) how do i get involved in
5138        pushing it forward
5139 
5140~master would nock be easier to understand of cores were introduced as
5141        "closures", or "objects"? or if gates were *just* functions and
5142        their nature as cores deferred or elided?
5143 
5144~widbur last i checked, 1 & 2 were undecidable and the answer to 3 was
5145        'we don't yet want people to get involved'
5146 
5147        to me, yes
5148 
5149        absolutely
5150 
5151        and if the networking were explained in terms of, say, standard
5152        consensus algorithms
5153 
5154~master as far as i know, the networking isn't really explained
5155        anywhere. but i've been out of the docs loop
5156 
5157        the network implementation is extensively and mysteriously
5158        explained
5159 
5160~widbur last time i asked about it, i was told to shut up and wait until
5161        it was rewritten
5162 
5163~master ;)
5164 
5165        well, it's been rewritten
5166 
5167~widbur responsiveness is much better
5168 
5169~master not sure if ~rovnys is in here
5170 
5171~mocrux In regards to naming, I think most programmers should be used to
5172        both name overloading and concept overloading (diff names for
5173        same concept) if they've jumped around enough across languages
5174        in genreal
5175 
5176~sarpen new ames is well documented
5177 
5178~widbur my cli doesn't work for chat, at all, but the web interface does
5179        (and i hate using browsers for cli things)
5180 
5181~sarpen (in the code)
5182 
5183~master is there a github issue for your cli problems?
5184 
5185        there have been some bugs
5186 
5187        you could |unlink %chat-cli and then |link %chat-cli from :dojo
5188 
5189        and i think there's a command to re-establish subscriptions
5190 
5191~widbur ~master, i don't know enough to know if they are bugs
5192 
5193~master what are the symptoms?
5194 
5195        i'm using the cli rn
5196 
5197~widbur i switch to chat-cli with ^x, ;join ~dopzod/urbit-help
5198 
5199        what i'm expecting at this point is a chat-cli
5200        ~dopzod/urbit-help prompt, and eventually to receive backscroll
5201 
5202        i get a chat-cli prompt, but no 'channel' prompt, no backscroll
5203        and no new messages either
5204 
5205        web interface works
5206 
5207~master from :chat-cli try: ;what
5208 
5209        or i guess ;chats, which is new
5210 
5211~widbur already have, it reports ~dopzod/urbit-help
5212 
5213        interestingly, when i try to ;join a channel i can't, like
5214        ~dalten/collective, i get the permission errors on the
5215        read/write but get the channel prompt
5216 
5217        chats reports this channel and hoontards
5218 
5219~master that's probably the channel settings on ~dalten, but the
5220        permissions stuff is new and i'm not sure
5221 
5222        does ;what show you glyphs before the channel names?
5223 
5224~widbur i vaguely remember that i needed to set the audience to write
5225        to a specific channel, but i should be receiving from all the
5226        channels i'm on, right?
5227 
5228~master that's right
5229 
5230        ;$glyph sets your audience on cli
5231 
5232        for me, this channel is &, so ;&
5233 
5234        and & is then at the end of my prompt
5235 
5236~widbur i just assumed ~dalten/collective doesn't exist because it
5237        was set up for the initial meeting only, or is locked down to
5238        participants only
5239 
5240        it would then make sense that i can't open its /read and /write
5241 
5242        urbit-help is @ for me, and @ is at the end of my prompt
5243 
5244~master i gotta step away in a minute, but i'll swing back around
5245 
5246~widbur no worries
5247 
5248~master if you type and a send a message, is it delivered to urbit-help?
5249 
5250        one more thing: from :dojo, try :chat-cli %connect
5251 
5252        i see your message in urbit-help
5253 
5254~widbur no, it isn't
5255 
5256        %connect returns >=, which i'm assuming is success
5257 
5258        i'm getting regular errors about eth watcher still running and
5259        so forth, i'm assuming that's normal
5260 
5261~master that command was a "poke" to the chat-cli app. pokes are rpc
5262        messages, positively or negatively acknowledged, but receiving
5263        no response
5264 
5265        so, yes
5266 
5267~widbur along with intermittent zod not responding, which i remember
5268        from last time, although they're less common now
5269 
5270~hodler i fixed my chat-cli with this
5271        https://github.com/urbit/urbit/issues/2099#issuecomment-56629214
5272        7
5273 
5274~master fyi, ~holder, urls are best as separate messages
5275 
5276        they can then be auto-linked, properly formatted on cli, etc
5277 
5278        simple messsage types, avoids markup overhead
5279 
5280~widbur ford is the http subsystem, isn't it? don't see how that would
5281        fix my cli
5282 
5283~master ~widbur, have you seen a chat cli output yet?
5284 
5285~widbur but willing to try
5286 
5287~hodler it worked for me
5288 
5289        my chat-cli wasn't showing anything
5290 
5291~widbur ~master, nope
5292 
5293~master ford is the build system, a la henry
5294 
5295        that linked command forces a rebuild of all apps
5296 
5297~widbur ah
5298 
5299~master needed to manually resolve some recently introduced bugs that
5300        broke the reactively of "live" builds
5301 
5302        just the :goad part
5303 
5304        the |wipe is to clear the %ford cache
5305 
5306~widbur makes sense
5307 
5308        didn't fix anything, but it does make sense ;)
5309 
5310        btw, urbit doesn't handle paths to ships correctly, which breaks
5311        the aur package and i'm assuming all other binary distributions
5312 
5313        only thing that would work for me is downloading the tarball,
5314        changing to that working dir and creating a comet in it
5315 
5316~master the filesystem path to the ship directory?
5317 
5318        or the path to the binary?
5319 
5320~widbur both
5321 
5322        looks like urbit expects urbit-worker to be in cwd
5323 
5324        and specifying the ship with a path doesn't work either
5325 
5326        so i can't cd-to-urbit-install-dir and specify
5327        /home/paul/mycomet
5328 
5329~master you should be able to specify the pier in that manner
5330 
5331        there's a couple issues with the binary path
5332 
5333~widbur and i can't cd to ~/mycomet and run urbit either via $PATH or
5334        via explicit path
5335 
5336~master our x-compilation of ncurses is just wrong, so we ship a
5337        terminfo database, which is horrifying
5338 
5339        local relative paths to the binary should always work
5340 
5341~widbur none of those combinations work as of right now with the aur
5342        package which is at the same version as the official download
5343 
5344~master back later
5345 
5346~widbur cheers
5347 
5348        terminfo isn't the problem, not finding urbit-worker or a
5349        spectacular sigsegv are ;)
5350 
5351        ~marzod, for when you're back, write via chat-cli seems to be
5352        working, no backscroll or read or prompt still
5353 
5354~somlus widbur still there?
5355 
5356~widbur soz, wish there were a good way to alert on new messages
5357 
5358        yes
5359 
5360        ~somlus, also soz for polluting your teaching channel with
5361        off-topic stuffs
5362 
5363~master hey ~widbur
5364 
5365        did your cli chat recover?
5366 
5367~somlus no worries wilbur all goo. was going to ask - do you need a
5368        planet?
5369 
5370        *good
5371 
5372        *widbur
5373 
5374~ronreg ~master-morzod: it actually is kind of confusing knowing that
5375        a gate (i..e the thing that looks like a function) is just a
5376        special core
5377 
5378        that cores are the most fundamental hoon data structure
5379 
5380~widbur ~master, no. writes work, reads, backscroll and prompt doesn't.
5381        i was going to 'turn it off, turn it back on' with the whole
5382        ship, or creating another comet, but i didn't want to do that
5383        unless you don't care about debugging this particular problem
5384        beyond making it work for me
5385 
5386        ~somlus, i'd like one in an ideal world, yeah, last time i
5387        checked in on urbit there wasn't a way for me to get one i
5388        think. there's a meetup in my area at the end of the month which
5389        ~winter is organizing and he was going to give me one there i
5390        think, but it's not a given that i'll be able to make that date
5391        because of time conflicts
5392 
5393        i'm not sure what the new system of sponsorship means, or what
5394        the etiquette around all of it is, in the sense that last time
5395        i checked the star->planet thing was meant to be a purely
5396        mechanical thing, whereas now it may or may not be a social
5397        construct of some kind?
5398 
5399~somlus if you email me your email address at somlus@somlus.com I will
5400        send you a planett.
5401 
5402        there is no significant social construct at all yet. I watch
5403        for people who look like they will contribute and I give them
5404        planets. You can check out:
5405 
5406        ~somlus-savlev/welcome
5407 
5408        ~nibfeb/welcome
5409 
5410        for more info. I'm headed to sleep fairly soon but if I see
5411        an email requesting a planet and providing an email address
5412        I'll send you an invite that includes a planet. you will be
5413        under my star nibfeb but here are no obligations and you can
5414        escape to another star if my star ever goes dark, which would be
5415        unintended.
5416 
5417~rabsef I think it remains to be seen what sort of service stars will
5418        perform for their planets. I don't think it's just the routing
5419        tho - or I should say I don't think Tlon expects it to be just
5420        the routing
5421 
5422~widbur future is hard to predict, but as of right now, the community is
5423        small enough that even if it's not meant as a social construct,
5424        it can't help but be one to an extent
5425 
5426~ronreg I've had conversations with people where they've said they want
5427        a star to be like an ISP
5428 
5429~widbur i've read that. every single isp relationship i've ever had has
5430        been abusive
5431 
5432~ronreg and that goes along with the "network infrastructure not social
5433        infrastructure" I've seen mentioned, and also the deliberate
5434        obfustication of what star a planet belongs to
5435 
5436        on the other hand, some people do talk as if a star should do
5437        things for planets that make being a member of a particular star
5438        socially valuable
5439 
5440        which makes stars less pure network infrastructure
5441 
5442~widbur people are social animals and most tend to want small in-group
5443        relationships at least
5444 
5445~ronreg yeah, but not necessarily with the same people who are in your
5446        star :)
5447 
5448~widbur among other things. sigils are meant to be random and not
5449        particularly interesting in terms of one vs another, but i've
5450        noticed more appealing ones being listed for higher prices
5451 
5452        indeed. once distinct group functionality is added as i've seen
5453        mentioned somewhere, it might take that over
5454 
5455        on the other hand, currently running a star 24/7 is a favour to
5456        planet owners and it's reasonable that there's some expectation
5457        of social reciprocity
5458 
5459        and repeated validation of reciprocity is how a high trust
5460        social group is built
5461 
5462~somlus good stuff, and what we should be talking about
5463 
5464~rabsef apropos of nothing I'm almost done w/ my card transformations
5465 
5466~widbur talking about it is fun, but i expect the result to emerge on
5467        its own regardless of discussion
5468 
5469        and what happens now, with a small community where over time
5470        most people will have talked to each other at length, and what
5471        happens later with more adoption will likely be very different
5472 
5473        every chance the folks who were early form an elite like the
5474        Old Bolsheviks and proceed to have champagne parties in large
5475        ballrooms, figuratively
5476 
5477~ronreg well that's an unsettling metaphor
5478 
5479        given that stalin killed many of them
5480 
5481~widbur .. then someone with a spectacular moustache sends us all
5482        somewhere cold to chop wood
5483 
5484        to be fair, this analogy could be past usefulness at this point.
5485        i can't decide who would be Kamenev for one
5486 
5487        ah, the multi-minute lag, old friend, how i've missed you ;)
5488 
5489        if the MSM is to be believed about Curtis, we should recast this
5490        into an analogy about Roehm and the SA? ;)
5491 
5492        personally, one of the things i'm super interested in seeing is
5493        how sigils end up playing out
5494 
5495~ronreg the MSM doesn't know who curtis is. it's specialty tech media
5496        who says he's a nazi
5497 
5498~widbur people usually want to, and are used to, selecting their
5499        pseudonyms. getting a random one is interesting, and could work
5500        for most people most times, but there are syllables being used
5501        that could be highly aesthetically unpleasant for particular
5502        people dependent on personal preference and native language.
5503        onomatopoeia works both ways
5504 
5505        i'm pretty sure i saw some reference to him in one
5506        pseudo-intellectual pseudo-left rag or another
5507 
5508        atlantic, slate, that sort of thing
5509 
5510        tech media is bad as well, i remember when they were all
5511        collectively trying to work out how much they could attack thiel
5512        for his private political views and donations given that he has
5513        some degree of magic intersectional immunity due to being a bit
5514        oscar wilde
5515 
5516        vast swathes of 'journalism' is unreadable to anyone with
5517        half a brain these days because it offends intelligence to an
5518        unbearable degree
5519 
5520~rabsef Yussss - just finished the triple cut of the solitaire cipher.
5521        Now only the counting cut and the keystream generation function
5522        to go.
5523 
5524~somlus Sir ~widbur, you are considered a good egg because you
5525        intelligently assess and desire the improvement of systems and
5526        you are respectful. Welcome aboard!
5527 
5528        haclex-moppyx! I like your name. welcome to the world...
5529 
5530        if you absolutely hate that name you can give the invite to
5531        someone else and we can try again but I quite like it.
5532 
5533        Oh and ~rabsef and ~haclex, FYI the CEO to Tlon, ravmel-ropdyl
5534        has stated from time to time on urbit-help that he wants that
5535        chat to be primarily about urbit help. Love the banter, but
5536        if you dive in to a convo on something, consider taking it to
5537        another channel. Feel free to use this channel if you like,
5538        though I have been considering having a nibfeb channel for some
5539        time so I might create one for all conversations that start on
5540        UH but then need be moved because of a non "urbit help" focus.
5541 
5542        Actually I just created it, feel free to continue the
5543        conversation there:
5544 
5545        ~nibfeb/uh-off-topic
5546 
5547~widbur cheers mate and yeah, point well taken on not going off-topic
5548        for too long, i would find that annoying myself to be fair
5549 
5550~somlus are you cool with the name? you are haclex-moppyx now.
5551 
5552        I think it's awesome, personally.
5553 
5554        It's very rare to have both a cool name and cool sigil BTW.
5555 
5556~widbur yeah, i'm well happy with that
5557 
5558        just concentrating now on getting through the process without
5559        losing keys and things
5560 
5561        you would not believe the amount of crypto i've got in wallets i
5562        will never be able to access again in all likelihood
5563 
5564        i don't actually know in fact, i've forgotten that as well,
5565        could be 2 shekels
5566 
5567        my brain doesn't cooperate with me on remembering things i
5568        consider important, it retains what it thinks is worth retaining
5569        and fks me off
5570 
5571~somlus haha! just make sure you download and store your master ticket
5572        and you'll be just fine. ok headed back to bed. midnight San
5573        Francisco time, cheers.
5574 
5575~widbur cheers mate
5576 
5577~ronreg I'm pretty sure I've never lost any cryptocurrency
5578 
5579~rabsef I was in a bitcoin mining collective early on and maybe got a
5580        half a bitcoin reward and then just lost the laptop that had the
5581        wallet so that’s cool
5582 
5583        also yah somlus good reminder
5584 
5585        joining ot
5586 
5587~widbur .
5588 
5589~somlus received
5590 
5591~widbur so i'm able to write here from my comet, but not join from the
5592        planet? this "maintenance-free operating function" thing is
5593        working out brilliantly ;)
5594 
5595~rapfyr maintenance-free is a goal not a reality.
5596 
5597~minder Archer -> Target
5598 
5599        We are currently the arrow.
5600 
5601~rapfyr which may or may not have tnt strapped to it :-p
5602 
5603~widbur i get that, but curtis claimed that making urbit a pure function
5604        gives you that for free
5605 
5606        like exactly-once delivery and so on
5607 
5608        it's more gentle ribbing than anything else
5609 
5610        last time i checked out urbit, it had the old chat mechanism and
5611        it worked perfectly
5612 
5613        the networking was dog slow and completely unreliable, but you
5614        could join channels, read the backscroll and write to them
5615        (eventually) all while receiving updates on the welfare of ~zod
5616        every second or so ;)
5617 
5618~master ~widbur, does this count as a sockpuppet attack?
5619 
5620        ;)
5621 
5622        exactly-once message delivery doesn't save you from
5623        implementation bugs, like delivering the wrong message
5624 
5625        unfortunately
5626 
5627~widbur it's only a sockpuppet if you can't see the hand and don't know
5628        to whom it belongs ;)
5629 
5630        this is the channel my ~haclex can't join
5631 
5632        although it's an interesting question socially what the
5633        etiquette ends up being around using more than one identity
5634 
5635~rabsef no etiquette. I'm actually you and ~master and this is all
5636        staged for ~rapfyr and ~somlus, who are both the same perosn
5637 
5638        person*
5639 
5640~widbur heh
5641 
5642        if you were me, you'd be busy screaming in horror, not posting
5643        here ;)
5644 
5645~rabsef exactly what we would say in this revealing situation
5646 
5647~minder ?
5648 
5649~widbur are those eyes or breasts? my font isn't specific. also, how did
5650        you do that?
5651 
5652~rabsef lol we see what we want to see comet
5653 
5654~widbur yeah, because breasts are impossible to find elsewhere on the
5655        internet
5656 
5657~rabsef fsir
5658 
5659        loool
5660 
5661        fair*
5662 
5663        ? <- wondering if all emoji work - thinking maybe cuz their utf
5664        8 right?
5665 
5666        apparently - that looks right to me
5667 
5668~widbur ah, right, i keep forgetting about that abomination of an abuse
5669        ;)
5670 
5671~somlus ~widbur why aren't you using your planet now?
5672 
5673~borlyt need some help with recursion, i get an error on the $(5 5) line
5674 
5675~rabsef I think the tape type can take utf8 so I assume this means
5676        you can put emoji in your programs in Hoon lol. this has not
5677        occurred to me until now
5678 
5679        be prepared for my next gen project which will only output in
5680        eggplant/peach emoji Morse code
5681 
5682        i think someone needs to explain how recursion works in hoon
5683        because i think i'm completely misunderstanding the workings of
5684        |- and $
5685 
5686~rabsef whatcha tryin to do there? my simplistic way of thinking of
5687        it is $ allows you to change the start conditions and start
5688        again at |- so $(a 5, b 5) would work there and exit with a= 5
5689        printing just 5
5690 
5691        you have to tell it which face ur modifyibf
5692 
5693        separated by commas
5694 
5695        the thing I think is least helpful in your example is that it
5696        only causes one recursion. you’re just giving the face a new
5697        value instead of modifying and recurring multiple times. you
5698        should try just $(a +(a)) which will end printing 10. you could
5699        also add a =| c=(list @ud). 1, change your finish condition to
5700        [a c] and your $ to $(a +(a), c [a c]) and see what a looks like
5701        each recursion
5702 
5703        typo - do not include “. 1”
5704 
5705        I am starting to understand it better. suggest walking thru
5706        the Fibonacci example too. %= $ is just a tall version of $()
5707        for our limited purposes so you can see more easily face ->
5708        transformed value -> recursion
5709 
5710        in my suggested construction your output should be like [10 ~[9
5711        8 7 6 5 4 3 2 1]]
5712 
5713        note that the $ transforms happen “simultaneously” or in other
5714        words +a is not represented in the first term of c there
5715 
5716        ~somlus - I think comet was having issues getting planet into
5717        this room... :(
5718 
5719~borlyt yeah i messed up with my $() syntax
5720 
5721~rabsef any luck there ~borlyt?
5722 
5723~walden hiya
5724 
5725        weird, sorry
5726 
5727        set: + ~somlus-savlev/hoontards
5728 
5729~widbur ~somlus, i am, but it can't join this channel. there's a bug, an
5730        issue has been filed, someone is purportedly working on it
5731 
5732        .
5733 
5734        .
5735 
5736~donbec hi
5737 
5738        hi
5739 
5740~sanlyt test
5741 
5742~somlus received.
5743 
5744        ah got it ~widbur. FYI I have seen that when I join channels
5745        too soon after launching a ship for the first time. It needs to
5746        equilibrate or update or something first. So now I launch new
5747        ships (or moons etc) and then walk away for 15-20 mins and then
5748        join everything. I have no idea if this is a real workaround or
5749        I just get lucky when I do it that way
5750 
5751        Oh also - my ship was offline for a bit today.
5752 
5753        my machine with the ubuntu install on it (where it lives)
5754        rebooted itself. I have to figure out if that was an
5755        update-related reboot or if I have an unstable system. This is
5756        the first instance
5757 
5758~hadpun ~somlus i do the same and it kinda works, ships (especially
5759        moons) need a while to set everything up, especially network
5760        wise
5761 
5762~somlus yeah moons are a mixed bag.. Sometimes they work great and
5763        sometimes they just barely hobble along. Sometimes they are
5764        birthed with some but not all of the planet's subscriptions
5765        automatically.
5766 
5767~hadpun quite weird
5768 
5769        when booting this one, i wasn't in any chats, but whatever my
5770        planet typed, regardless of chat, I could see
5771 
5772~somlus That's odd. I think that they are focusing their energy on
5773        non-moon features and fixes, which makes sense when they are
5774        trying to make Urbit ready for the public this yeary.
5775 
5776~hadpun do you by chance know if windows support is on the way? i have a
5777        bunch of friends who would join pretty fast if they didn't have
5778        to buy a mac/install linux
5779 
5780        yeah
5781 
5782        can someone explain what the "subject" is
5783 
5784        been going through this rn
5785 
5786        https://urbit.org/docs/tutorials/hoon/the-subject-and-its-legs/
5787 
5788        and am quite confused
5789 
5790~rabsef the subject is dependent o what sort of thing you are running.
5791        in a “naked gen” the subject is the entirety of the generator
5792        inclusive of the sample aka arguments
5793 
5794        in other contexts it can be the entiret of the operating
5795        environment
5796 
5797~somlus hadpun windows support is not coming but a coming windows update
5798        will allow for an actual real deal linux kernel in the ubuntu
5799        install so the ubuntu VM on windows should run flawlessly and
5800        urbit can be installed on that. for now you can easily install
5801        ubuntu on the free version of VMWare
5802 
5803~rabsef generallt the subject is “everything”. imagine it like this
5804        - your brain runs programs on your experience and knowledge,
5805        and any new incoming information becomes part of that set of
5806        knowledge/ experience - further and finally, your brain programs
5807        can be operated against as well qua subject - in the case of say
5808        evaluating your thinking process
5809 
5810        another possible example. everything in urbit is a noun
5811        including programs the biggest noun, is the noun of all
5812        operational data available aka everything
5813 
5814~ronreg in a naked gen the dojo environment has to be part of the
5815        subject too, right? otherwise you wouldn't be able to call std
5816        library functions
5817 
5818~rabsef and that noun is the subject - kinda
5819 
5820        I think that’s partly true. Hoon.hoon is available but now eny
5821        beak are not and they’re part of dono
5822 
5823        maybe I’m wrong ~ronreg. great point
5824 
5825        one can also call custom libraries in a naked or other gen afaik
5826        so only the base case is “just Hoon.hoon” if that’s right, at
5827        all.
5828 
5829        is it just me or is this the most inefficient way to implement
5830        addition
5831 
5832        addition is a simple bit operation, isn't it?
5833 
5834        this is in the hoon.hoon file
5835 
5836~rabsef When I don't understand an implementation, I say "Oh that's
5837        because Nock works that way" - how does that feel as an answer?
5838        ;)
5839 
5840~hadpun kind of helps, but when it comes down to machine operations this
5841        just seems slow af
5842 
5843        imagine (add 1000 1000)
5844 
5845        you'd need 1000 actions
5846 
5847        whereas just adding the 2 rows of bits is quite easy to do for
5848        hardware
5849 
5850        idk maybe i'm thinking of it wrong
5851 
5852~rabsef That's the concept of a jet
5853 
5854~hadpun a jet?
5855 
5856~rabsef nah I think you're right - that's why a lot of mathematical
5857        actions are jetted - but the idea is that Nock has to be able to
5858        perform them too
5859 
5860        a jet is an outside assisted action
5861 
5862        doesn't perform in nock
5863 
5864        as long as jets are trusted, then there is no problem
5865 
5866        https://urbit.org/docs/tutorials/vere/jetting/
5867 
5868~borlyt so it's like for example C-code that's being run instead of
5869        nock?
5870 
5871~rabsef A "jet" is a sort of software device driver in Urbit. Urbit is
5872        precisely defined in terms of an interpreter that fits on a
5873        T-shirt, Nock. But a naive implementation of Nock is nowhere
5874        near practical, because Nock has only one integer operation:
5875        increment. For example, to decrement a number in Nock, you
5876        increment a counter up to the number under it.
5877 
5878        They agree w/ you
5879 
5880        correct
5881 
5882        https://urbit.org/blog/common-objections-to-urbit/#jets
5883 
5884        also jet just sounds cool
5885 
5886        tbh
5887 
5888~borlyt that seems like a massive flaw in nock, no?
5889 
5890        they're using hotfixes to make up for bad decisions made in
5891        the writing of nock, isn't this the problem Unix has but to an
5892        extreme degree? isn't urbit supposed to fix this?
5893 
5894        the point of the rewrite of our entire OS is to get a clean
5895        slate and "do it right this time"
5896 
5897~rabsef You should read that objection
5898 
5899~minder +1 ~rabsef's statement
5900 
5901~borlyt I get the general idea they're going for but addition really
5902        shouldn't be this hard. it's literally 1 hardware operation
5903 
5904~mocrux It just depends on the base primitives you work from.
5905 
5906        If you choose to use a very very simple system from the core
5907        (you encode less inherent primitives) well then you need to
5908        define things like this.
5909 
5910~rabsef I think for further elucidation you should ask someone from Tlon
5911        about it because I lack the requisite understanding to defend it
5912        properly
5913 
5914        but they're all in a big fun meeting together so it might be a
5915        min
5916 
5917~mocrux It's kind of like looking at System F/lambda calculus and saying
5918        it's too complicated to do anything with.
5919 
5920~sarpen it's a common paradigm in many functional languages, say you're
5921        dealing with natural numbers, they're recursively defined as
5922        nat=zero or successor(nat)
5923 
5924        and then operations are defined in terms comparing how many
5925        succ there are in a value, for type checking to work out
5926        appropriately, but then they'll do something analogous to
5927        jetting for actual computation
5928 
5929        ^
5930 
5931        it's like looking at the natural numbers and saying "we need
5932        the reals" without going through and proving the reals can be
5933        constructed from the nats
5934 
5935~mocrux The tradeoff is you create a system that makes sense and is
5936        analysable from bottom-to-top for a human inspecting it, but
5937        that you can fit in black boxes that speed things up without
5938        having to needlessly look into the black boxes and adding
5939        essentially useless complexity.
5940 
5941~widbur no one says the implementation has to be naive
5942 
5943        it just has to be functionally (punz ftw) equivalent
5944 
5945        i'm not at all sure that side-effects and ephemeral data are a
5946        problem for OSes, and therefore that their removal is going to
5947        be brilliant
5948 
5949        it's useful in pure functional languages because it is a way
5950        to control and limit shared state, thereby reducing perceived
5951        complexity from a given vantage point
5952 
5953~ronreg ~sarpen-laplux: this is precisely how Agda works
5954 
5955        you define the naturals in a provably sound way, and then add a
5956        pragma annotaiton to tell teh compiler to map these to machine
5957        integers
5958 
5959        same with booleans and a few other types
5960 
5961        that said, I don't htink agda is a terribly practical language
5962 
5963        and I think "The fact remains, however, that a Nock interpreter
5964        without jets is completely useless. And there is not even an
5965        informal document that tells you what jets you need for a
5966        reasonably efficient Urbit. (There should be such a document.)"
5967        matters quite a bit
5968 
5969~mocrux ~ronreg, I think it's common among most if not all the
5970        dependently typed languages/theorem provers that allow for
5971        executable code.
5972 
5973        Though I'd agree documentation is definitely important for
5974        explaining what are the bare essentials for jets
5975 
5976~rabsef Anyone understand %~ and want to help me understand how ~(. door
5977        arg) can be used to then run the arms of that modified door?
5978 
5979~sarpen ~(arm core sam) resolves core, replaces it's sample with sam,
5980        and then evaluates arm with respect to it
5981 
5982        with arm=. that means the entire core
5983 
5984        so it's technically not always the arm, but arm-style wing paths
5985 
5986        (I think)
5987 
5988        one would use ~(. door arg) or door(arg) mostly when I think you
5989        want to replace the sample of a door, and then call arms from it
5990        later
5991 
5992        call/read
5993 
5994~rabsef but what I want to do is once having modified the door/core,
5995        call it w/ the arms to modify against the sample and then get
5996        the modified sample back
5997 
5998        Yes - I think you're right
5999 
6000        ~(. cardmanipulation swapdeck) in my case gives me the arms in
6001        my door like <5.has> just like if you .add in dojo
6002 
6003        what good does that do me tho? How do I use that?
6004 
6005~sarpen oh you want an arm to return a modified version of the door it's
6006        in? i.e. its parent's core?
6007 
6008        an arm that's a gate*
6009 
6010~rabsef how do you do that last bit
6011 
6012        >call arms from it later
6013 
6014        Not quite - I'm not explaining it well
6015 
6016        I have a core that has an arm that is a door. That door/arm
6017        manages a deck of cards
6018 
6019        It has a bunt value of x deck. Each arm inside that door
6020        modifies the deck in certain ways
6021 
6022        What I need to be able to do is call that door several times,
6023        use the arms against the bunt value of the deck (or subsequent
6024        decks that came out of the last time the door and arms were
6025        called) and return the deck
6026 
6027        https://pastebin.com/DhWyU9b0
6028 
6029        see near bottom for my door
6030 
6031        I think maybe the problem is that my door's arms aren't gates so
6032        they don't generate output - they just manipulate the sample
6033 
6034        so ~(arm door arg) doesn't work cuz there is no output
6035 
6036        and since each arm operates against the sample again, I can't
6037        nest them like (arm door (arm door (arm door arg)))
6038 
6039        does that make sense?
6040 
6041        I could rewrite the arms to have outputs but i just was
6042        wondering if there's a way to proceed from here
6043 
6044~sarpen ~(arm door arg) should work just fine
6045 
6046        if it's a gate, ~(arm door arg) would receive the gate - not yet
6047        evaluated
6048 
6049~rabsef but ~(arm door arg) doesn't produce anything - how do I request
6050        arg back from ~(arm door arg)
6051 
6052        it's not a gate - that's my issue, I think
6053 
6054        ti's just some hoon that operates against initial arg or
6055        supplemented arg
6056 
6057~sarpen I think the problem might be that cardmanipulation is a gate
6058 
6059        er not a gate
6060 
6061        but a core
6062 
6063        >but ~(arm door arg) doesn't produce anything
6064 
6065        it produces nothing at all? or
6066 
6067~rabsef I might be able to write everything down there again as gates
6068        but I was hoping not to lol
6069 
6070~sarpen no you shouldn't have to
6071 
6072~rabsef sorry - I'm getting a little lag and not seeing your comments
6073 
6074~sarpen if the only arg you need is in the door it shouldn't be a
6075        problem
6076 
6077~rabsef ~(. cardmanipulation swapdeck) gives me the number of arms and
6078        other information a la .add in dojo
6079 
6080~sarpen right
6081 
6082~rabsef ~(jokerafunc cardmanipulation swapdeck) I think gives me a nest
6083        fail
6084 
6085        let me try again
6086 
6087        If I just put ~(joker card swap) after giving swap a face at the
6088        top I get a nest fail
6089 
6090        I can try =/ test ~(joker card swap) and then outputting test
6091 
6092        nest fail as well
6093 
6094        My supposition is that I am nest failing because joker doesn't
6095        produce a gate that would take the doors arm and do something
6096 
6097        instead it just modifies the arg (arg not arm above) for the
6098        next arm of the door. If I just do =~(cardmanipulation arm arm
6099        arm arm arm currentdeck) I get the current deck that I want to
6100        us eas the starting deck next go round
6101 
6102        bUt I don't understand how to pass that and then call the door
6103        again
6104 
6105        I tried =~((cardmanip swapdeck) arm arm2 arm3 armn current deck)
6106        but it didn't like that
6107 
6108        ere =~(cardmanip(currentdeck swapdeck).....
6109 
6110        Imagine this simple example from bank account
6111        https://urbit.org/docs/tutorials/hoon/bank-account/ <- how would
6112        you recurse that so the second time you called it you started
6113        with acct balance = 150
6114 
6115        assuming the door existed in a larger core, nd each time you
6116        recur you have to use the value to do some other operation in an
6117        arm of the parent core
6118 
6119        and maybe this is the wrong strategy for doing that - so i'm
6120        open to ideas
6121 
6122        I'm just going to rewrite them as gates and nested call them as
6123        described
6124 
6125~sarpen >~(jokerafunc cardmanipulation swapdeck) I think gives me a nest
6126        fail
6127 
6128        hmm
6129 
6130        lemme reread in morn
6131 
6132~rabsef I rewrote it
6133 
6134        and it still doesn't quite work - output wont cast into a
6135        meaningful form
6136 
6137~sarpen might try to hack on it
6138 
6139~rabsef https://pastebin.com/AVCHVZeE
6140 
6141        now it's all gates but when I call even one of them the output
6142        is a mess of shite
6143 
6144        can't cast it into a meaningful form
6145 
6146~sarpen maybe try =+ cm=~(. cardmanipulation swapdeck) jokerafunc:cm
6147 
6148        it shouldn't be gates
6149 
6150        where are you trying to call it?
6151 
6152~rabsef at %~(jokerafunc cardmanipulation swapdeck)
6153 
6154        trying that w/ the other copy - one min
6155 
6156        nest fail
6157 
6158~somlus sorry about the lag, folks. I might have to breach if I can't
6159        figure this out. I might change my hardware location to one
6160        where I get fantastic up/down speeds. At home (where it lives
6161        now) I have 1GbPS down but only 40 Mbps up.
6162 
6163        ~rabsef what are we covering tomorrow. Also in my effort to
6164        figure out why things are so laggy, can you tell me how many
6165        (more or less) chat channels and notebooks you are subscribed
6166        to? I am subscribed to about 35 chats and 20 or so notebooks.
6167        Shouldn't be that, but I'm looking at all angles now.
6168 
6169~rabsef tomorrow we will slavishly hack away at my generator until it
6170        works
6171 
6172        after going through
6173        https://urbit.org/docs/tutorials/hoon/recursion/ and
6174        https://urbit.org/docs/tutorials/hoon/recursion/
6175 
6176        we should review https://pastebin.com/ac3Yc1ky and comment it
6177 
6178~somlus rock it out. let's do it. I will work on that today.
6179 
6180        oh those two links separated by "and" were the same. is there
6181        another link to go through other than the recursion one?
6182 
6183~rabsef https://urbit.org/docs/tutorials/hoon/fibonacci/
6184 
6185        my b
6186 
6187~somlus got it
6188 
6189        Interesting, I unsubscribed from some chats and now it's
6190        responsive as hell. We'll see how long this holds up
6191 
6192~rabsef yah it's faster here too
6193 
6194~somlus I only unsub'd to like 3-4
6195 
6196~rabsef I'm in like 10-15 and subscribed to like 8 notebooks
6197 
6198~somlus one of them was your snapple one!
6199 
6200~rabsef to answer your prior question
6201 
6202~somlus thx
6203 
6204~rabsef lol oh well that doesn't exist rn so that might be the problem
6205 
6206        oh no wait it does
6207 
6208        ~rabsef-bicrym/snapplefans
6209 
6210~somlus yeah, it actually might be a problem to be subscribed to ones
6211        that don't exist. Perhaps it goes on fetching and fetching.
6212 
6213        wow i'm like in a little heaven here of responsiveness for a
6214        while.
6215 
6216        I'd be pretty surprised it it was my internet speeds TBH. must
6217        be something else.
6218 
6219~rabsef yah uve never been this snappy here
6220 
6221        tbh - is nice
6222 
6223~somlus wow. I wonder if that is it
6224 
6225~rabsef I'm going to wrastle this thing to the ground today
6226 
6227        by hook or bloody crook
6228 
6229~somlus I admire your tenaciousness
6230 
6231~rabsef I just took the goddamned door out and I think it's gonna work
6232 
6233        I wish I understood doors but I'll take working over that
6234 
6235        fudging awesome
6236 
6237        I just found a previously unnoticed bug in one of my arms now
6238        that they're written as gates
6239 
6240        I think I'm there.
6241 
6242        it works
6243 
6244        I did it
6245 
6246~somlus what was the bug? just something tiny?
6247 
6248~rabsef 1 math issue cuz I'm dumb
6249 
6250        then I just decided to go back to my original idea of using a
6251        counter declared to do a transformation twice the first time and
6252        once every subsequent time
6253 
6254        it works - posting soon
6255 
6256        https://pastebin.com/pk0RLwdW
6257 
6258        actually I think I don't need the counter
6259 
6260        Doing the transformation for tempvalcard twice always works
6261 
6262        I forgot to even increment passone
6263 
6264        that... doesn't make sense
6265 
6266        oh yes it does I just still have trouble w/ recursion occurring
6267        simultaneously
6268 
6269        hot damn
6270 
6271        I'm projecting one into the future each time and adding the past
6272        value at the same time
6273 
6274        WILD STUFF MAN I'm so pleased
6275 
6276~somlus kapow!!!
6277 
6278~rabsef Probably don't need to triple post this but here we go:
6279 
6280        I done it.
6281 
6282~borlyt Caesar cypher?
6283 
6284~somlus Here is my commented hoon 101 3a exercise. See you at 9am (1:13
6285        from now)
6286 
6287        https://pastebin.com/DqyWFNzu
6288 
6289        Will be stepping away, but a couple of questions confuse me and
6290        i did not include them in the paste.bin because it would get too
6291        busy. Q1: so it looks like recursion includes the line where $
6292        lives, and once that line's code is complete THEN it jumps back
6293        to the trap beginning?
6294 
6295        Q2: It's interesting that a variable can name itself in the line
6296        of code whereby the variable assumes a new value. I just want
6297        to be sure that I understand that the nested code is completed
6298        first (before the variable changes value) and then the outer
6299        code happens whereby the variable assumes the new value. Seems
6300        like that must be it.
6301 
6302        Headed on a short errand, should be back in time
6303 
6304~rabsef ~borlyt - solitaire cipher. can be decoded with a physical pack
6305        of cards
6306 
6307        https://en.m.wikipedia.org/wiki/Solitaire_(cipher)
6308 
6309        or encoded for thst matter - gonna build in custom deck starting
6310        state functionality next
6311 
6312        Hello ~somlus!
6313 
6314~somlus Hello sensei
6315 
6316~rabsef :-D Are you ready to begin?
6317 
6318~somlus Very shortly
6319 
6320~panlux Talking as moon
6321 
6322~rabsef reading you loud and clear
6323 
6324~panlux opening pandora
6325 
6326        opening multiple windows on multiple screens (believing,
6327        falsely, that flair equals competence)
6328 
6329~rabsef It definitely contributes to it
6330 
6331~panlux ready, boss. Do you have the pastebin open for my commenting on
6332        that code?
6333 
6334        and there are two questions above
6335 
6336~rabsef I do and I read
6337 
6338~panlux 3 2 1 go
6339 
6340~rabsef Let's start w/ your pastebin - There's one or two things I'd
6341        like to clarify/modify in how you thought through this
6342 
6343~mocrux That's why I always have a script that just prints random files
6344        of code fullscreen at blazing fast speed on a 2nd monitor to
6345        make it feel like I'm in a hacker movie 24/7 while coding
6346 
6347~panlux yes!
6348 
6349~rabsef 1) Line 4 - >(that is this program) - > Are you stating that the
6350        sample is the program? Is that right?
6351 
6352~panlux you honed right in on something I was lazy about. I need to get
6353        the vocab right. generators, samples, parts of generators called
6354        wings and limbs. Still don't have it all down, so I say program
6355 
6356        here I meant this entire set of code
6357 
6358        all falls under |=
6359 
6360~rabsef Let's just review the terms briefly. A gate (|=) consists of
6361        a battery and payload. The battery is the one arm named $,
6362        the payload is the data required to compute $. The payload is
6363        further divided into sample and context. The Sample, rightly, is
6364        the input value or /n/ in our case
6365 
6366~panlux thank you that is clear. However, I thought that $ was specific
6367        to the recursion, and was only a marker / trigger for recursion.
6368        Here it seems to have a much more important definition
6369 
6370~rabsef second, @ud is not technically "real numbers" Both @s and @u
6371        would be real numbers or integers. @ud is an unsigned decimal
6372        which could be real.
6373 
6374~panlux should have looked that up instead of assuming
6375 
6376        I see it now on my printout list of auras
6377 
6378~rabsef So that's a little past my understanding but the way I think of
6379        it presently is that $ is the arm of computation that occurs to
6380        produce an output. This could happen through recursion where $
6381        could be explicitly called and modified several times or, in the
6382        case of a trivial gate like |= a=@ud a where input is printed
6383        back, I think it is still the unnamed $ that we are computing
6384 
6385        or the undeclared $ i should say
6386 
6387        uncalled* even more rightly in line w/ the prior paragraph
6388 
6389~panlux is $ specific to the rune |= ?
6390 
6391        or is it used all over the place?
6392 
6393~rabsef It is used in all cores, as far as I know. But, in gates, there
6394        is only one arm and it is named $. In higher level cores, there
6395        can be other named arms in addition that contribute to the
6396        computation of $
6397 
6398~panlux actually if $ is "the thing we are computing" then that actually
6399        clarifies things for me. I have often been lost in the code
6400        as to exactly where the output is stored. In Basic (I know, I
6401        said basic, it's from my youth) you were only naming values to
6402        variables, and in excel you are always outputting to a cell.
6403        So here I had a hard time telling exactly what the output is
6404        outputting to
6405 
6406~rabsef For instance, in my pontifex gen, there are like 10 arms, but
6407        the final computation is in $ and reads from the other arms to
6408        do things
6409 
6410~panlux ok good that does clarify things
6411 
6412        awesome. I'm surprised that is not front and center in the early
6413        part of the docs.
6414 
6415~rabsef yah $ is like the cell or the value that you print in basic
6416 
6417~panlux boom. hard to believe but that is a major epiphany for me.
6418 
6419~rabsef awesome!
6420 
6421~panlux we are seeking $
6422 
6423        always (or mostly always so far)
6424 
6425~rabsef there's some nuance in that comparison insofar as like the
6426        cell can be recursively re-computed before output, right, and
6427        technically the output is the result of computing $ but whatever
6428 
6429~panlux olk
6430 
6431        ok
6432 
6433~rabsef ok, so you formed the trap yep, then let's look @ line 16
6434 
6435        Yes, at first t is 1 but we only know t is 1 if n is 1 first
6436        iteration, right?
6437 
6438        t /starts/ as 1 but $ will modify t based on some parameter to a
6439        new value each recursion
6440 
6441~panlux oh yeah, we are counting down so would this be the stop point
6442        then?
6443 
6444        I thought it was an early stop in case the user entered 1, but
6445        it works either way as n counts down to 1. stops there with an
6446        accurate t value
6447 
6448~rabsef an important note here - failing to modify the values you expect
6449        to modify in $ can easily result in an endless loop - imagine if
6450        we never decremented n and gave it +thiscore 3
6451 
6452        it would just multiply 3 * t endlessly
6453 
6454~panlux got it. does the system hang then or know there is an error and
6455        tell you?
6456 
6457~rabsef It is /both/. If the user enters 1, the recursion is never
6458        called and the trap never returned to
6459 
6460        the output is t, which by default is 1
6461 
6462        the system spins / until you ctrl-c basically
6463 
6464        if you do it bad enough it crashes - depends on what you're
6465        making the poor thing do
6466 
6467~panlux ok
6468 
6469~rabsef ok so let's look @ 19. I would like it if you thought of 19 as
6470        including the idea "if n!=1 then..."
6471 
6472        Because ?: has two potential outcomes, just like
6473        =IF(condition,true,false) in excel
6474 
6475        I assume you thought this implicitly based on 16 but it's good
6476        to remember it
6477 
6478        also this helps remind us of explicit children in hoon
6479 
6480        and you have the function of the $ down correctly. In other
6481        words, you might say - reiterate the loop w/ these changed
6482        starting conditions
6483 
6484        So - I think unless it would be dull to you, we might think
6485        through the products and give ourselves some example starting
6486        n's?
6487 
6488        Incidentally, if you've not seen that notation before != means
6489        not equal to
6490 
6491~panlux ah I did not get it. you give me too much credit. So this line
6492        is the "then" line if n does not equal one correct?
6493 
6494        yeah was just about to ask
6495 
6496        got it
6497 
6498~rabsef Correct. As I too easily brushed over - $ qua starting again w/
6499        changed initial values does not occur if n=1 at the first pass.
6500        Instead $ is computed as t
6501 
6502~panlux got it. What is "- $ qua" in your most recent statement?
6503 
6504        I mean "$ qua"
6505 
6506~rabsef qua = "defined as " or "in the sense of"
6507 
6508~panlux thank you
6509 
6510~rabsef sorry. So let's talk through this gate as if we are it.
6511 
6512        Our user has just given us a /sample/ of 2
6513 
6514~panlux great
6515 
6516~rabsef What is our first step? I will give it to us - first, we are
6517        given as known that t=@ud = 1
6518 
6519        What happens next?
6520 
6521~panlux it states that n is that number that was given to us
6522 
6523~rabsef sorry - yes technically the first step is definitely that we
6524        give a name or face to 2 that is n
6525 
6526        very good.
6527 
6528~panlux then the next step to inform hoon that there is a trap that is
6529        being set up and sets up the start point
6530 
6531        and right after that it says is the number the person entered (2
6532        in this case) equal to 1?
6533 
6534~rabsef Yes - so we know we may be coming back here again if we are told
6535        to
6536 
6537        Yes - and is it?
6538 
6539        lol
6540 
6541~panlux If yes (though it's not), then output t, but since it is not
6542 
6543        it goes to two lines down which is code line 20
6544 
6545        and then in 20, I do want to get into the weeds here.
6546 
6547~rabsef Ok so may I? or would you like to describe?
6548 
6549~panlux I'll try
6550 
6551        it says "ok so here is the recursion point - be advised Hoon,
6552        after we're done with this line, go back to |-"
6553 
6554~rabsef Yes
6555 
6556~panlux and then it says, ok I'm going to redefine n, but first I will
6557        use n in it's OWN redefinition. NEW n is going to be equal to
6558        CURRENT n minus 1. But we will just call it n when this is done
6559 
6560~rabsef yes - all redefinition 'kicks' or 'takes effect' post $ line
6561 
6562~panlux and it also says similarly, I am going to redefine t, but first
6563        use it also. I want to multiply t by NEW n (but now just called
6564        n).
6565 
6566~rabsef no
6567 
6568        that is the one error in thinking - and I have it too
6569 
6570~panlux ok that was a question
6571 
6572~rabsef yes sorry
6573 
6574~panlux no worries at all, just commenting.
6575 
6576        it reads from right to left
6577 
6578        ?
6579 
6580~rabsef I do that exact thing still tho. Just wanted to let you know
6581        that it's hard to get over
6582 
6583        lol
6584 
6585~panlux so the t calculation happens with the old n? and then the old n
6586        is used once again to make the new n?
6587 
6588~rabsef No - not that either. Instead, what happens in $() only takes
6589        effect thereafter
6590 
6591~panlux wow. ok
6592 
6593~rabsef So $ says "For the next pass, starting from trap, please use n
6594        as n-1 and t = as (current n * t)"
6595 
6596        but current n is still the n to wit in n-1
6597 
6598~panlux and also the mul t n
6599 
6600        ok got it. so you have to think of this line as needing to be
6601        complete and the very finishing touch is to make the changes on
6602        the variables before Hoon leaves this line
6603 
6604~rabsef yes that's what I mean - so with our values, we would say "Start
6605        again at trap w/ n = 1 and t = 2 * 1
6606 
6607~panlux ok cool
6608 
6609~rabsef yes - I like that - the finishing touch is to take effect
6610 
6611        and that happens while we the program are cycling back up to 11
6612 
6613        (this one goes to 11)
6614 
6615        so now we are at 11 w/ n=1 and t=(2*1) or t=2
6616 
6617~panlux ok cool! I just ran the code and it works. 5 produced 120 and 10
6618        produced 3.629.800
6619 
6620        mf year
6621 
6622        yeah
6623 
6624        (nice spinal tap reference)
6625 
6626~rabsef lol yah so what's another name for this function?
6627 
6628~panlux factorial
6629 
6630~rabsef hell yah
6631 
6632        So before we look @ something a little harder - any questions?
6633 
6634~panlux sorry i jumped ahead for fun. we should finish out what happens.
6635        Then with n=1 when you get to the logic question, n=1 and so it
6636        spits out t, and that is 2
6637 
6638        yes one question
6639 
6640~rabsef ^correct re: final step
6641 
6642~panlux You said that the program produces $, but it actually does not -
6643        it produces t and that is spit out at the end
6644 
6645        so let's briefly revisit what you mean by it has $ as a final
6646        product
6647 
6648~rabsef Yes - this is the nuance $ is the computation that leads us to a
6649        leg. A leg is anything that is not an arm, i.e. a fully computed
6650        noun
6651 
6652        $ is an arm, i.e. something that still requires computation
6653 
6654        once buc is settled it is the output (t)
6655 
6656~panlux is t the leg? or what would be the leg here?
6657 
6658~rabsef t is a leg, yes. Buc acts on the leg t until the arm collapses
6659        to no further computations
6660 
6661        it also acts on teh incoming leg n
6662 
6663        and it does so in everything from 11 to 20
6664 
6665        Maybe think of $ as the current operation at the (usually) top
6666        of the program - that which will collapse
6667 
6668~panlux why is n a leg? is user input considered a process with an
6669        endpoint?
6670 
6671~rabsef a user input that is a noun that is not a computation is a leg
6672 
6673~panlux ok I do think it is useful to think of it that way. $ is
6674        the actual meat of the computation. the rest is sort of
6675        decision-making and establishing rules
6676 
6677~rabsef if the input were like a gate somehow it would be different.
6678 
6679        Here is a forinstance - in my pontifex deal, I have some nested
6680        arms - so you do one operation to the deck then another (arm
6681        (arm (arm starting leg)))
6682 
6683~panlux actually it's very useful to think of it that way. ANother
6684        imporant addition to the hoon intro section. (we need to
6685        first create rules and then flow and then ultimately make a
6686        computation. ok let's do that)
6687 
6688~rabsef something like that - the leg is a leg because it doesn't
6689        require any computation. Before the arms are computed inside
6690        to outside, they are arms - as they collapse into values they
6691        create legs for the next arm
6692 
6693~panlux copy that on user imput
6694 
6695~rabsef $ technically (I think) includes the decision making as that is
6696        part of what tells us /how/ to computate
6697 
6698        does that track for you?
6699 
6700~panlux I think so yes
6701 
6702        So since I have made two generators now, I need to call my
6703        mother and tell her I am a coder.
6704 
6705~rabsef lol do it
6706 
6707~panlux (opens window with falling matrix symbol animations)
6708 
6709        ok so what's next chief?
6710 
6711        we have a good 20 mins
6712 
6713~panlux https://www.youtube.com/watch?v=rpWrtXyEAN0[embed]
6714 
6715        matrix raining code for 1 hour
6716 
6717~rabsef https://urbit.org/docs/tutorials/hoon/fibonacci/
6718 
6719        We're going to look @ that.
6720 
6721        Need a few initial data points
6722 
6723~panlux ok greqat
6724 
6725~rabsef =+ declares lots of variables together at once (rightly faces
6726        for nouns)
6727 
6728        %- does something to the output thereafter itself
6729 
6730~panlux by line:
6731 
6732        https://pastebin.com/U4LA1QPt
6733 
6734~rabsef %= resolves a "wing" (arm) based on a series of changes
6735 
6736        Sorry - yes. So 2 is operate "flop" (invert what we have)
6737        against our output, 3 is decalre many variables (could use =/
6738        repeatedly) and 6 is just a very long $(blah)
6739 
6740        /in this case/ - %= could modify some other wing, but for us
6741        here it is buc
6742 
6743        oh - one more is the part after the trap on line 4 - it just
6744        casts (or prints) our output as in the form of r as declared (a
6745        list of @uds)
6746 
6747~panlux ok cool, digging deeper
6748 
6749~rabsef I wouldn't worry to omuch about those specific parts - but I'll
6750        fill in gaps if we get stuck walking through.
6751 
6752~panlux sounds good.
6753 
6754~rabsef I would suggest you get a paper to write some values on - cuz
6755        we're going to want to do n = 3 or something to see what happens
6756 
6757~panlux yes, I actually print the code in very large print and write all
6758        over it. That is how I am getting through this.
6759 
6760        Another thing I will sugges to Tlon
6761 
6762~rabsef One additional trick they're using here - remember what I said
6763        about avoiding infinite loops and making sure to operate on your
6764        variables? Here in %= we do not change n! How is this possible?
6765 
6766~panlux will take me some time to get to that
6767 
6768~rabsef look @ line 5 and line 7
6769 
6770        what are we doing
6771 
6772        note the starting value of i at line 3
6773 
6774        I will tell you if you would prefer - lmk
6775 
6776~panlux it is zero
6777 
6778        actually I think we are too far along for me to follow. Better
6779        to begin at the top and go through what each line is doing, more
6780        slowly.
6781 
6782~rabsef Yes ok. 1 - we ask for input of an @ud
6783 
6784        ignore 2 - we will come back to it
6785 
6786~panlux ok got it
6787 
6788~rabsef 3 - we declare 4 faces, i p q and r
6789 
6790~panlux in three, =+ means combine a new noun with the subject. ok we
6791        are defining variables, but I get stuck at r
6792 
6793~rabsef i = 0, p = 0, q = 1, r is a blank (*) or bunt value (don't worry
6794        about it - blank) list of @uds
6795 
6796~panlux 'cause it's a weirdo
6797 
6798        ok will nto worry about r for now
6799 
6800~rabsef Let's talk about a list briefly. A blank list is [~]
6801 
6802        Oh I just meant don't worry about * or bunt. We need to worry
6803        about r
6804 
6805        lol
6806 
6807        a list is [a [series [of [nested [cells [terminated [by ~]]]]]]]
6808 
6809~panlux ok
6810 
6811        got all of that
6812 
6813~rabsef The computer politely prints it as [a series of nested cells
6814        terminated by ~]
6815 
6816        but we must remember we are creating cells each time., read from
6817        right to left
6818 
6819~panlux how cordial! And I came empty-handed.
6820 
6821~rabsef yes well it's used to it
6822 
6823~panlux yes, right to left
6824 
6825        haha
6826 
6827~rabsef what is (this is just another notation - putting the ~ out
6828        front) ~[1 2 3] in full write out?
6829 
6830~panlux it is [1 [2 3]]
6831 
6832~rabsef not quite
6833 
6834        do not forget "terminated by ~"
6835 
6836~panlux ~[1 [2 3]]
6837 
6838~rabsef one might also write it [1 2 3 ~]
6839 
6840        it's just different notation - annoying but you must get used to
6841        it
6842 
6843~panlux ok got it
6844 
6845~rabsef ~[1 2 3] is [1 [2 [3 ~]]]
6846 
6847        /terminated/ by ~ always in a list
6848 
6849        e.g the last term is always ~
6850 
6851        ok so we know our variables. Let's move to 4, set a trap, ignore
6852        the cast, and move to 5
6853 
6854        What happens on 5
6855 
6856~panlux it says...
6857 
6858~rabsef Here is a hint: Hoon does not know the difference between two
6859        spaces and a carriage return
6860 
6861~panlux no it does not
6862 
6863        ah thanks
6864 
6865~rabsef they are synonymous
6866 
6867~panlux it's a logic test. if i and n are equivalent, then r is the
6868        final output. if not then go to 6 and continue
6869 
6870~rabsef precisely
6871 
6872~panlux got me with those two spaces
6873 
6874~rabsef now let's look @ line 7, assuming for the moment that it is part
6875        of a $() progression
6876 
6877~panlux ok
6878 
6879~rabsef what do we do to i=0 on each recursion
6880 
6881        (just i rather, but starting at =0)
6882 
6883~panlux add one to it
6884 
6885~rabsef +(x) == (add 1 x)
6886 
6887~panlux (sorry was looking something up)
6888 
6889~rabsef yes - just making sure you knew that notation
6890 
6891~panlux yes got it
6892 
6893        on line 7 we will add 1 to i at the conclusion of the $ section
6894 
6895~rabsef OK - so to return to my question earlier, how can we get away w/
6896        not modifying n as we did in factorial?
6897 
6898        In factorial we counted n down to 0. Here, we count 0 i) up to n
6899 
6900~panlux we are modifying other numbers instead and then comparing them
6901        to n - the opposite solution
6902 
6903~rabsef (i)*
6904 
6905~panlux counting up not down
6906 
6907        oh sorry we are not adding one to i I read that wrong
6908 
6909~rabsef precisely
6910 
6911        no you're reading that right
6912 
6913        each recursion we will add (increment up) one to i
6914 
6915        increment is the only mathematical operator that knock has (to
6916        point to our discussion of jets a few days back)
6917 
6918        nock*
6919 
6920~panlux so +(2) would be 3?>
6921 
6922        3?
6923 
6924        +() is the add one function?
6925 
6926~rabsef ok - so back at 5 for a moment - when i == n, we output r. Else
6927        we go to 6 and do some things
6928 
6929        yes +(2) == 3
6930 
6931~panlux yeah
6932 
6933~rabsef ok - so the things we do - we know line 7, what is line 8?
6934 
6935~panlux cool, so we add one to i, then we define p as q? (not sure why
6936        yet)
6937 
6938~rabsef yes - you will see. and 9?
6939 
6940~panlux add em together and that results in what will be the NEW q at
6941        the end of the $ sequence
6942 
6943~rabsef and do note that "add them together" means add the /old/ p
6944        despite the fact that p has just been modified for purposes of
6945        the next pass
6946 
6947~panlux yes, thank you for that reminder. that is important to note
6948 
6949~rabsef same for line 10 then - what happens?
6950 
6951~panlux line 10 I do not understand cince there is no command like "add"
6952 
6953~rabsef great
6954 
6955~panlux I don't know what putting numbers in a list does other than
6956        output that same exact list
6957 
6958~rabsef this is shorthand and annoying also
6959 
6960~panlux oh making a longer list?
6961 
6962~rabsef this is the rune :-(a b) which forms a cell of [a b]
6963 
6964~panlux oh thank you.
6965 
6966        man we need them runes but I do see it on the hoon cheat sheet
6967 
6968~rabsef so, to give it away, each recursion we will take r (which starts
6969        as [~]) and make a larger cell, the first of which will be [q ~]
6970        (~[1 2 3] = [1 [2 [3 ~]]] - see what we are doing)
6971 
6972        we are hacking together a list in the long notation
6973 
6974~panlux ok so if you start with list r such as [1 2 3 4] and q is 13 you
6975        get [13 1 2 3 4]
6976 
6977        reading yourse
6978 
6979~rabsef yes yes but do not forget the ~
6980 
6981        r will always have ~ at the terminus
6982 
6983~panlux true, also we are building to the right, mine built to the left
6984 
6985~rabsef its "bunt" value is ~, we do not even need to tell the computer
6986 
6987~panlux if the code was r [r q] then would it build to the left?
6988 
6989~rabsef no we are building to the left. "adding to the head"
6990 
6991~panlux got it, I am understanding that i need to use ~ and not get
6992        sloppy on that
6993 
6994        ok always add to the head.
6995 
6996~rabsef let me show you - if R already equals [1 2 ~] then the next pass
6997        we will make [q [1 2 ~]]
6998 
6999~panlux ok I'm finally understanding much of this
7000 
7001~rabsef because [x y] == :-(value value)
7002 
7003~panlux copy that
7004 
7005~rabsef and remember that a list is a series of nested cells - which is
7006        exactly what we are making
7007 
7008        it is "computationally costly" to build a list on the right hand
7009        side
7010 
7011        because effectively we must change EVERY cell, not just add one
7012        cell to the exterior
7013 
7014        does that follow?
7015 
7016~panlux makes sense, yes I understand.
7017 
7018~rabsef that is adding 3 to the end of [1 [2 ~]] makes [1[2[3~]]]
7019 
7020~panlux It's like taking one single base pair out of genetic code - has
7021        massive downstream implications
7022 
7023        and tis tis just says go back to the recursion line 4?
7024 
7025~rabsef sorry I forgot spaces there lol
7026 
7027        tistis closes centis
7028 
7029        centis has an indeterminate number of children
7030 
7031~panlux oh ok
7032 
7033        ah, got it. ok finally this makes sense.
7034 
7035~rabsef however centis is modifying buc and buc will send us back to the
7036        trap
7037 
7038~panlux at least something needs closure!
7039 
7040~rabsef lol yah
7041 
7042        I do on many of my past relationships! oh that's not what you
7043        meant huh
7044 
7045~panlux haha! no closure on the other code was very hard for me.
7046 
7047~rabsef ok so we have most of this - let's just hack thru it w/ n=3 so
7048        we have to do it 3 times
7049 
7050~panlux ok so == closes this little redfining group of instructions and
7051        then we have reached the end so the recursion goes back to 4..
7052 
7053~rabsef starting at line 5. 0!=3 so not r but...
7054 
7055~panlux ok
7056 
7057        increase 0 to 1
7058 
7059~rabsef tell me what each variable in turn will be changed to
7060 
7061        yes
7062 
7063~panlux p is 0 and will become q (after ==)
7064 
7065        q is 1
7066 
7067        q will become the addition of old p and old p, which will be 1
7068 
7069~rabsef old p and old q I'm sure you meant.
7070 
7071~panlux yes sorry
7072 
7073~rabsef np just making sure.
7074 
7075        and what of r
7076 
7077~panlux then r will be the list generated by old q (1) and a currently
7078        empty r, only definied by the stuff that will go into it but not
7079        yet by contents
7080 
7081~rabsef what will it look like
7082 
7083        if we outputted r immediately after this what would r b
7084 
7085~panlux and the real change then is that p has become 1 and r will
7086        become ~[1]
7087 
7088        ?
7089 
7090        can you have a list of one atom?
7091 
7092~rabsef p is not involved with r. R will become [1 ~] by proxy of q
7093        starting at value 1
7094 
7095        yes, you can have a list of 1 atom - it is just a trivial list
7096 
7097~panlux wait, just summarizing the changes at the conclusion of $, p
7098        does become r, so p goes from 0 to r (1)
7099 
7100        oops shit
7101 
7102        q
7103 
7104~rabsef ok - so to define our recursion conditions -> i=1, p=1, q=1,
7105        r=[1 ~]
7106 
7107        and n still equals 3
7108 
7109~panlux yeah let me regroup on your observation and drop my line of
7110        thought as it was messed up
7111 
7112        yes so all of that looks good, and then we go again through the
7113        $ instructions
7114 
7115        i WILL become 2
7116 
7117~rabsef ok because 1 !=3, yes - go on
7118 
7119~panlux p WILL not change, or rather it will once again be set to q
7120        which is 1
7121 
7122~rabsef yep
7123 
7124~panlux q WILL become the sum of p and q which WILL be 2
7125 
7126~rabsef write the long form cell of r
7127 
7128~panlux and then r WILL become the former list plus the value of q in
7129        the head or ~[1] WILL become ~[2 1]
7130 
7131~rabsef not quite
7132 
7133~panlux I don't think there is a longer form of that
7134 
7135~rabsef think about that again
7136 
7137~panlux oh
7138 
7139~rabsef and write it with all the [[[[]]]]s
7140 
7141~panlux oh
7142 
7143        [2 [1 ~]]
7144 
7145~rabsef Hint: r is currently [1 ~]
7146 
7147        bingo
7148 
7149        didn't need a hint
7150 
7151        no lol
7152 
7153        wait
7154 
7155        why 2?
7156 
7157~panlux let's see
7158 
7159~rabsef 2 is incorrect but your form is right
7160 
7161        q old q is 1, q becomes 2 when? after ==
7162 
7163        [1 [1 ~]]
7164 
7165~panlux true. I didn't think there were two 1's in the solution but
7166        there it is
7167 
7168        you're right
7169 
7170~rabsef ok, so now we are back at ?: =(2 3)
7171 
7172        thus, not r, so...
7173 
7174~panlux back to recursion point with new variable values, i=2 p=1 and
7175        q=2
7176 
7177~rabsef I suggest you write all the changes in one line so you do not
7178        think of them as occurring in order
7179 
7180        yes that is right and that takes us to 6 which is where I meant
7181        my last comment to apply
7182 
7183~panlux yeah, that is where I get screwed up. I will do that as an
7184        exercise. I will do this one by hand soon
7185 
7186~rabsef Do me it on this recursion then - starting conditions i=2 p=1
7187        q=2 r=[1 [1 ~]]
7188 
7189~panlux ah ok this is easy i is counting the number of recursions
7190 
7191~rabsef instead of n counting down
7192 
7193~panlux up to n, the total amount of recursions we have asked this thing
7194        to do
7195 
7196~rabsef yep
7197 
7198~panlux ok so with those starting conditions, i WILL become 3, p WILL
7199        become 2, and q WILL become 3, while r gets an old q (2) added
7200        to its head, which will make it [2 [1 [1 ~]]]
7201 
7202        and we then can restart with i=3 p=2 q=3 r is what I put there
7203 
7204        I'll be damned
7205 
7206~rabsef right?
7207 
7208        it starts to click after a while
7209 
7210~panlux and it keeps going until we have it does the number of times we
7211        have told it to repeat. this is useful because we can easily
7212        make a loop of n times now with this simple toolkit
7213 
7214~rabsef we are not quite done
7215 
7216        2 things remain
7217 
7218~panlux oh I was going to make n-5 chocolate souffles
7219 
7220        n=5
7221 
7222~rabsef currently we are outputting r because i=n now
7223 
7224~panlux ok what remains?
7225 
7226~rabsef r is [2 [1 [1 ~]]]
7227 
7228~panlux ah yes. once i=n , then spit out r and for godsake let the
7229        little Hoon rest, will ya? (last part implied)
7230 
7231~rabsef but that is not a list - that is a series of nested cells (which
7232        technically is what a list is but we want to pretty print it
7233        with the ~ out front
7234 
7235~panlux ok
7236 
7237~rabsef so - we have ^+ r which "casts" r as a list
7238 
7239        take my word for that part
7240 
7241        then we have line 2
7242 
7243~panlux ok will do
7244 
7245        oh and flop is reverse it?
7246 
7247~rabsef which takes our list ~[2 1 1] and does something
7248 
7249        yes
7250 
7251~panlux I'm amazed that flop is up there at the top. why is that, when
7252        we read left to right and top to bottom and this work is done?
7253 
7254~rabsef because flop nests around 1 argument
7255 
7256        that argument is the computation of $
7257 
7258~panlux oh the rest is a dependency
7259 
7260~rabsef 1 child flop has only one child, and we must make the arm a leg
7261        for it to work
7262 
7263~panlux a child of flop
7264 
7265~rabsef (arm (arm (arm starting condition)))
7266 
7267        exactly
7268 
7269~panlux I hope my daughter is not a child of flop.
7270 
7271        ok my simple homework is going to be to make a list that omits
7272        the duplicate 1. Will be an easy chance so I will do that and
7273        get back to you
7274 
7275~rabsef lol
7276 
7277        flop does work at least.
7278 
7279~panlux interstingly, though, this flip completely deranges the meaning
7280        of the original list.
7281 
7282        i wonder when that would be used
7283 
7284~rabsef technically 1 is not duplicated - that's the nature of the
7285        fibonnaci sequence
7286 
7287        fibonacci*
7288 
7289        it's not wrong w/ the 2 1s that is
7290 
7291        nah fibonacci sequence is 1 1 2 3 5 etc
7292 
7293~panlux ah yes I recall.
7294 
7295        it's been a while since I studied that sequence. just wanting to
7296        do a simple exercise.
7297 
7298~rabsef You should write a recursion that does something else. Like
7299        makes a list of all n counting up to an input n where n is
7300        incremented once each recursion
7301 
7302~panlux ah yea, 1+1=2, 2+1=3, 3+2=5, etc
7303 
7304        yeah I will
7305 
7306~rabsef basically n output = n, but as a list showing how we got there
7307 
7308~panlux ok very fair
7309 
7310~rabsef so like a list of all natural numbers leading up to and
7311        including n that is
7312 
7313~panlux fantastic will do and then will have it count up to name
7314        ~somlus-savlev (not really)
7315 
7316~rabsef lol u could do that
7317 
7318        it'd just take a long time
7319 
7320~panlux cool man this as always has been fun and fruitful! I will review
7321        all of this stuff this week and look forward to our new topics
7322        next Sunday.
7323 
7324~rabsef I'll think about where we can go next
7325 
7326        it was a delight :)
7327 
7328        talk to you soon ~somlus
7329 
7330~somlus me too. back to my somlus visage.
7331 
7332        talk soon, and thanks. Very fun times.
7333 
7334~rabsef just built in custom deck support for pontifex gen
7335 
7336~risruc awesome stuff ~rabsef!
7337 
7338~somlus what does pontifex gen mean?
7339 
7340        oh that's the card generator
7341 
7342~risruc pontifex is the name of the encryption mechanism created by
7343        pontifex in Cryptonomicon
7344 
7345        i suggested ~rabsef look at building it, since he already had
7346        the card selection mechanism built and was looking for a next
7347        project.
7348 
7349        and then the guy went ahead and did it.
7350 
7351        like a boss.
7352 
7353~rabsef it was a great challenge - thank you again!
7354 
7355~somlus oh yeah! Sorry I even looked at that wikipedia entry on it
7356 
7357~rabsef while driving today I realized there’s one code bug which I can
7358        fix, but what that means to me is more than making a mistake. I
7359        can conceptualize in Hoon now. I can think about thinking about
7360        it
7361 
7362~somlus Very nice! And apparently you can put on after shave as a
7363        handsome bald man...
7364 
7365~rabsef ^this msg took a while to get delivered
7366 
7367        it's the feelsgoodman face! or at least the one that isn't a
7368        pepe
7369 
7370~somlus ah boy maybe not fixed after all, but at least a little better.
7371        tomorrow will move ships to a non comcast ISP and then we will
7372        see
7373 
7374~rabsef u gonna put it on the lil boy comp again
7375 
7376        I can only imagine laptop turns off a lot
7377 
7378~somlus I think I have the system settings set up properly to never
7379        sleep or turn off. my main concern is that the NUC has a bios
7380        setting to always restart after a power failure, while the
7381        laptop does not have such a setting. So the NUC is a bitter
7382        computer to recover from power outages on its own - just need
7383        scripts that auto-start my ships on boot.
7384 
7385~rapfyr that is awesome ~rabsef
7386 
7387~rabsef Thanks ~rapfyr; feels good to hear that too :)
7388 
7389~mocrux I'll be looking forward to the "Novel Hoon Design Patterns #X"
7390        notebook entries you'll be running with in a few months ~rabsef
7391 
7392        Assuming you keep up with this pace ;)
7393 
7394~rabsef That is really encouraging and I am not going to let you down,
7395        at least insofar as keeping up the pace - Doubt I'll have novel
7396        contributions too soon, but I hope to be a (some kind of)
7397        contributor to this project in short order.
7398 
7399        Thank you.
7400 
7401~talfus :publish|subscribe ~master-lardyr %recettes-de-urbit
7402 
7403        :publish|subscribe ~middev %urbit-resources
7404 
7405        :publish|subscribe ~hatdyl-toldeg %eudaemonia
7406 
7407        :publish|subscribe ~master-lardyr %urbit-journal-club
7408 
7409        :publish &publish-action [%subscribe ~hardux-rigdel
7410        %self-reliance]
7411 
7412~somlus Those go in dojo, you are probably aware
7413 
7414        hit control X and then when you see your ship name with dojo,
7415        re-enter those
7416 
7417~rabsef test
7418 
7419        last test
7420 
7421        ~somlus - I've lost the name of our last chat due to breaching
7422        again - thought I had it in a screenshot - apparently I did not.
7423        Please remind me at your convenience
7424 
7425~hadpun does hoon have a sort of hashmap? key value pairs and stuff
7426 
7427~rabsef there’s a map logic section in the tutorials on urbit.org
7428 
7429        I might not be answering correctly but it has map - yes
7430 
7431        also see the caeser cipher
7432 
7433        https://github.com/rabsef-bicrym/urbitasofia
7434 
7435        If you need some other examples of map functions, check out
7436        Pontifex on GitHub :-D
7437 
7438        hoontards coming soon. we will cover recursion in more advanced
7439        forms today.
7440 
7441        back in 15 ish. ~somlus, let me know if you can’t make it
7442 
7443~somlus sorry!
7444 
7445        I messaged you on the DM and should have checked this one.
7446 
7447        I'll be here.
7448 
7449        firing up my moon now
7450 
7451~rabsef sorry - brt
7452 
7453        missed the dm. this thing needs notifications
7454 
7455~somlus no worries.
7456 
7457        I'm having a real problem with moon communciations but that's
7458        ok. I can code there and talk here.
7459 
7460~rabsef that should work fine
7461 
7462        ok I'm at computer - let me boot my ssh connection and we'l lget
7463        started
7464 
7465~somlus sounds great
7466 
7467        I wanted to pre-annotate your code but I did not. I did print it
7468        out for annotation here
7469 
7470~rabsef For anyone following along here and lurking, we're going to look
7471        @ https://pastebin.com/vK1DTLAS
7472 
7473        and the ackermann function in the tutorials
7474 
7475        Ok first things first, we're going to need to know a few things
7476        about lists that you might already know
7477 
7478        1) A "tape" is a list of cords. A cord is a unicode character
7479        surrounded in ''
7480 
7481~somlus ok got it
7482 
7483~rabsef so 'a' is a cord 'at' is a cord, a tape is ['a' [' ' ['b' ~]]
7484        and so on
7485 
7486        2) weld joins lists together first to second
7487 
7488        so weld list1 list2 becomes list1list2
7489 
7490~somlus copy that
7491 
7492~rabsef 3) cass takes a tape and makes it lower case, cuss takes a tape
7493        and maeks it upper case
7494 
7495        4) addressing the head or the tail of a list, requires first
7496        that you determine that it is non-null, and then allows you to
7497        do so via i.list or t.list respectively
7498 
7499        aight - that's all we really need to know to get started w/ the
7500        pastebin
7501 
7502        so we have some pretty basic text up to line 9 - we should be
7503        familiar w/ those parts by now
7504 
7505~somlus cool. that is very straightforward, though I'm sure we will go
7506        into i.list or t.list more later
7507 
7508~rabsef to review, we're simply declaring alternator as 1, declaring a
7509        blank outputstring and requesting (line 3) a valued incoming
7510        string
7511 
7512        yes actually, let's start here - in dojo create yourself a tape
7513        with a face
7514 
7515        =abc "a b c"
7516 
7517        if you "abc" in dojo from there, you'll get back "a b c"
7518 
7519~somlus done
7520 
7521        abc
7522 
7523        sorry
7524 
7525~rabsef lol np
7526 
7527        you're calling the face and getting the value
7528 
7529~somlus got it, but in quotes
7530 
7531        (oh as you said I would)
7532 
7533        got it.
7534 
7535~rabsef no quotes for what you put in dojo, sorr, so if you >abc in
7536        dojo, you get back "a b c"
7537 
7538        you get it
7539 
7540        anyway try i.abc which should give us a
7541 
7542        what do you get?
7543 
7544~somlus ?
7545 
7546        I get an error.
7547 
7548~rabsef not what you hoped for
7549 
7550        this is because you have to make the list a lest or a non-null
7551        list first
7552 
7553        How do we determine if a list is null? do you konw the rune?
7554 
7555~somlus https://pastebin.com/PsJicDiq
7556 
7557        ok
7558 
7559        ~ is null but I do not know the rune
7560 
7561~rabsef ?~ can check a list for 'null-ness' for us, and it has two
7562        children - a condition if true and a condition if false
7563 
7564~somlus ?~ make sense.
7565 
7566~rabsef Try ?~(abc %.y %.n)
7567 
7568        in dojo
7569 
7570        just like in excel, the children are in order true then false
7571 
7572~somlus %.n
7573 
7574        is what I get.
7575 
7576~rabsef right, because abc is non null
7577 
7578~somlus got it
7579 
7580~rabsef we're on course. Now, let's try i.xxx - ?~(abc "this won't
7581        happen" i.abc)
7582 
7583~somlus it did not accept that, stopped at -
7584 
7585~rabsef Sorry - didn't mean the - from here I'll put dojo code in curly
7586        brackets to distinguish
7587 
7588        exclude the curlies
7589 
7590        {?~(abc "x" i.abc)}
7591 
7592~somlus got it.
7593 
7594        similar error
7595 
7596        to the one above
7597 
7598~rabsef if you just {abc} do you still get your tape?
7599 
7600~somlus that I paste.bin'd
7601 
7602        interesting, no I do not
7603 
7604        same error
7605 
7606~rabsef let's set it again {=abc "a b c"}
7607 
7608        then repeat the above {?~(abc "x" i.abc)}
7609 
7610~somlus I'm an idiot. I'm still pasting in th {} chars
7611 
7612        will try again
7613 
7614~rabsef lol nbd
7615 
7616~somlus so {abc} works but
7617 
7618~rabsef does {abc} yield output of "a b c" in double quotes, to confirm?
7619 
7620~somlus yes sir but the {?~(abc "x" i.abc)} is not working
7621 
7622        (without {} of course)
7623 
7624~rabsef what does it do?
7625 
7626        same error?
7627 
7628        https://pastebin.com/5vg42Z0p
7629 
7630~somlus now it does work, even though I put in the same thing 'a' is
7631        returned
7632 
7633~rabsef ^what you should get
7634 
7635        ah yes good
7636 
7637        idk what happened time 1 but it doesn't matter
7638 
7639        you should get 'a'
7640 
7641~somlus got it
7642 
7643~rabsef the head of the list abc is the cord 'a'
7644 
7645        a tape is a list of cords
7646 
7647        let's clear that by simply {=abc} in dojo to remove the saved
7648        face abc and destroy our data
7649 
7650~somlus i.abc specified head?
7651 
7652~rabsef yes sorry i is head and t is tail
7653 
7654~somlus done, cleared
7655 
7656~rabsef if you want you could do with {=abc "a b c"} then {?~(abc "x"
7657        t.abc)} and get " b c"
7658 
7659~somlus (also incidentally this is all very well described, you are
7660        going to be GREAT at creating lessons)
7661 
7662~rabsef why double then and not ' '?
7663 
7664        let's actually do that
7665 
7666~somlus I got " b c:
7667 
7668~rabsef why?
7669 
7670~somlus sorry " at end
7671 
7672~rabsef why isn't it in single quotes?
7673 
7674        Hint: nested cells
7675 
7676~somlus is the answer something I've already learned? i honestly don't
7677        know though I do understand nested cells. I do not know how that
7678        relates to ' vs "
7679 
7680        BRB in 2 mins, a handyman doing an ethernet install wants to ask
7681        me something. my apologies
7682 
7683~rabsef I don't really expect you to intuit it but asking an impossible
7684        question might spur the answer to hae impact
7685 
7686        no problem - waiting - tell me when you're back
7687 
7688        Also, you could technically answer this given what we know from
7689        fibonacci but again I wouldn't expect it per se
7690 
7691        Answer: [the head [everything [else [is [tail ~]]]]]
7692 
7693~somlus i'm back
7694 
7695~rabsef The head of a list is the exterior +2 value (remember the noun
7696        is +1, the head is +2, everything else is +3
7697 
7698        so when we ask for i.abc, we get the value at +2 which is a
7699        cord, the tape is a list of such cords
7700 
7701~somlus ah. the tail for (a b c) is [b c]
7702 
7703~rabsef when we ask for the tail, however, we get back the list that
7704        remains because we are not asking for the cord at position
7705        +6 (in my bad example, the word everything), but instead the
7706        entirety of +3 (the list everything else is tail)
7707 
7708        precisely - although more specifically, the spaces are also
7709        cords
7710 
7711~somlus ah, so its [ b c]
7712 
7713~rabsef let's clear abc {=abc} and make a tape that begins with a space
7714        to prove it {=space " tape"}
7715 
7716        {?~(space "x" i.space)}
7717 
7718~somlus I did not register mentally that spaces are cords. that is
7719        important since in the [1 2 3] notation the spaces are only
7720        separators of numbers
7721 
7722~rabsef yes - that is confusing
7723 
7724        that's a good point
7725 
7726~somlus I got a fail for
7727 
7728        ?~(space "x" i.space)
7729 
7730        shit
7731 
7732~rabsef did you set space
7733 
7734~somlus didnt' do the space command first.
7735 
7736~rabsef sorry I am going all over the place a little
7737 
7738        perhaps too much coffee :)
7739 
7740~somlus ' '
7741 
7742~rabsef there's our proof
7743 
7744        let's unset space {=space}
7745 
7746~somlus no worries, it's fun. I like to need to keep but but be able to
7747        keep up (even if barely)
7748 
7749        done
7750 
7751~rabsef let's now return to the code
7752 
7753~somlus ok
7754 
7755~rabsef line 3 - ignore the first cell of now eny beak
7756 
7757        the second cell at line 3 requests a value w/ a face of
7758        incomingstring
7759 
7760        that value must be a tape
7761 
7762        so we will be giving this generator a tape
7763 
7764~somlus got it
7765 
7766        (let's take a break for just a moment and talk about
7767        instruction)
7768 
7769~rabsef we'll cover maybe later how a say generator works and what we
7770        do with those arguments, but for us now it is effectively |=
7771        incomingstring=tape
7772 
7773~somlus This is critical
7774 
7775~rabsef yes
7776 
7777        hit me
7778 
7779~somlus It's important to state at the very beginning in plain english
7780        "ok, so Hoon needs help at the start of every generator. since
7781        there are lots of ways of stating things, you must tell it
7782        exactly how you have inputted information."
7783 
7784~rabsef Yes - there is talk of naive type interpretation that isn't
7785        helpful
7786 
7787~somlus This is when we talk about how one thing "bead" or whatever, can
7788        mean multiple different things depending on the aura.
7789 
7790~rabsef I see what you mean
7791 
7792~somlus And the reason for this is the following:
7793 
7794~rabsef Potential Outline for Course Structure
7795 
7796        Programs we might be familiar w/ - input->output
7797 
7798        Types of input and output in hoon - introduce main concepts of
7799        atom, cell, list, and types of atoms/cells/lists (@ud, @tD, @p,
7800        cord, tape)
7801 
7802~somlus when someone looks at code, they should be first honing in on a
7803        few simple things. A beginner has no idea what to hone in on.
7804        But if we spoonfeed them and say "ok when you see code, the
7805        first few things you should do are...." and perhaps it is "look
7806        at the first generator" or "take a look at what sort of input is
7807        being requested" or something, then we can give them an actual
7808        process to begin instead of being lost.
7809 
7810~rabsef that's very good - I will add that as a note to the notes
7811 
7812~somlus So I think the lesson here is, it would be helpful to the
7813        students to provide some very early basic structure that they
7814        need to know for most programs of a given category, like "take
7815        user input and do something to it"
7816 
7817        I meant look at the first rune not generator
7818 
7819        yeah, input/output programs is a category. The course structure
7820        could go through like 2 different programs (easier and harder)
7821        for each "type" of program we decide is needed
7822 
7823        ok LMK if if you have thoughts on that, if not we can continue
7824        now that we have defined the user input as a tape and given it a
7825        variable name
7826 
7827~rabsef Yes - I think you're right. We can talk about input-> output
7828        then talk about types of input/output, then talk about the need
7829        to recognize type requirements at each level of a program (later
7830        you may have to recast things many times, or cast arm output for
7831        other arm use)
7832 
7833        the most important starting type check is what input is
7834        requested
7835 
7836~somlus and really emphasize it with annotations.
7837 
7838        ok cool
7839 
7840~rabsef So - taking that as granted for now, we should do that here.
7841        And perhaps we should just state outright the intention of this
7842        program
7843 
7844        our INPUT will be a tape "tape" and our OUTPUT will equally be a
7845        tape "TaPe" where we want to play w/ the capitalization (insert
7846        sarcastic spongebob meme)
7847 
7848~somlus good point^^ always state outright the intention of the program.
7849 
7850        awesome.
7851 
7852        AND
7853 
7854        in the instruction materials, probably add a step whereby you
7855        question "ok how do we think about doing this?"
7856 
7857        and the answer is "well, we have to go letter, by letter,
7858        obviously, so we have to create a generator that is able to
7859        slice off one letter at a time, and then alter every other one.
7860        for that, we need to be able to understand how Hoon thinks of a
7861        line of text and then based on that how we can chop it up and
7862        alter it, and then put it back together again"
7863 
7864~rabsef Yes. That is where I intended to send our minds using abc and
7865        space
7866 
7867        that's a good point - it should be said first then do those
7868        experiements to see the results
7869 
7870~somlus "said first concretely" "thinking about how we will go about
7871        this" "execution with lots of annotations & questions, answers"
7872 
7873        maybe. just thinking out loud
7874 
7875        I love the experiments BTW.
7876 
7877~rabsef So assume we talked through that and then did the abc and space
7878        experiments - we now have tools required to start down our road
7879        to our goal
7880 
7881        at least in part
7882 
7883~somlus because they both tell you what something is and what it isn't.
7884 
7885        and that creates mental bookends of boundaries around something
7886        strange which adds predictability and also a sense of
7887        safety/less confusion.
7888 
7889        yeah, this is great.
7890 
7891~rabsef so maybe we shouldn't have unset abc and space yet
7892 
7893        let's reset them {=abc "abc"} and (slightly modified {=space "
7894        def"}
7895 
7896~somlus ok...
7897 
7898        done
7899 
7900~rabsef Let's learn some basic "standard library" (pre programmed
7901        short-hands for other programs) tools we can use on our values
7902        named abc and space
7903 
7904        {(flop abc)}
7905 
7906~somlus as expected cba
7907 
7908~rabsef flop allows us to reverse the direction of a list - this is
7909        helpful in hoon because, given that lists are nested cells, it
7910        is easier to peel off the exterior layers
7911 
7912        as we peel off an exterior layer, apply a modification to it,
7913        and then do something with it, we will likely create our output
7914        in inverse order
7915 
7916        thus flop is important to set things back in the order we want
7917        after the fact
7918 
7919        let us do another (weld abc space)
7920 
7921        {(weld abc space)} rather
7922 
7923~somlus abc def
7924 
7925        are we able to do something like =weldedones (weld abc space)
7926        and define weldedones as that?
7927 
7928~rabsef right ok - so as we get lists we want to combine, we can combine
7929        them using weld, and if we find we are building our output in
7930        the wrong direction, we can correct with flop later
7931 
7932        we are
7933 
7934        let's try it
7935 
7936        and then {weldedones}
7937 
7938        (also looking @ DM - happy to talk by phone after - tell me when
7939        to call it)
7940 
7941~somlus yes
7942 
7943        cool ok
7944 
7945~rabsef (or in chat - w/e - just mean happy to do that)
7946 
7947        yah ok - so let's unset those values now, I think, and return to
7948        the code at hand
7949 
7950~somlus cool. let's get to whatever stopping point you see fit and then
7951        we can go to the chat channel
7952 
7953~rabsef {=abc} {=space} {=weldedones}
7954 
7955        it's important to clean up the dojo so you don't tack things as
7956        required into later programs
7957 
7958~somlus done
7959 
7960        cool
7961 
7962~rabsef whatever is saved in the dojo when you create a program is
7963        implicitly part of that program (weird, right?!)
7964 
7965        that is what people mean when they say that hoon is subject
7966        oriented
7967 
7968~somlus interesting, thank you for explaining that
7969 
7970~rabsef whatever is saved in dojo is implicitly part of the state, and
7971        therefore part of programs made thereafter
7972 
7973~somlus makes sense.
7974 
7975~rabsef ok - so we know that we should look @ input type when we analyze
7976        our programs - this program, on line 3, requests input of tape
7977 
7978        we can look @ line 6, also, to confirm that our output will be a
7979        tape also
7980 
7981        this coincides with our goal
7982 
7983        we have one hanging bit of information in line 5 that indicates
7984        that we are doing something with a number
7985 
7986        alternator is a number (@ud) that starts at 1
7987 
7988        we should keep those in mind when we evaluate the rest of the
7989        program - as these are likely the only things that will be acted
7990        on by our program
7991 
7992        not even likely, necessarily
7993 
7994        We know by now what a trap is on line 7
7995 
7996        so we know we will be doing something of recursion
7997 
7998~somlus i'm following you fine
7999 
8000~rabsef but what is different from our prior examples here is that,
8001        looking at lines 11 13 and 14 it appears there are multiple
8002        bucs?
8003 
8004        how can this be?
8005 
8006        Let's start walking through the trap
8007 
8008        Line 8 - We first do something we did earlier with abc and
8009        space. We ask if our incomingstring tape is null
8010 
8011        we know what we might be able to do with this. In this case, the
8012        TRUE value is to (flop) our output string
8013 
8014        as stated in prior lessons, we should not build a program that
8015        has an infinite recursion, so we know that, at some point,
8016        incomingstring /will/ be null, to allow us to proceed to output
8017 
8018        so - we should expect to find something that sends
8019        incomingstring to null-ness at some point - pin that thought
8020 
8021~somlus ok
8022 
8023~rabsef But, let's assume we don't pass a null string - so we'll ignore
8024        the line 9 outcome for the immediate present
8025 
8026        Line 10 - what does line 10 ask the computer to do?
8027 
8028~somlus branch on a logic test, and
8029 
8030~rabsef think of passing our old value space as incomingstring
8031 
8032        what is it checking
8033 
8034~somlus see if the head of the tape is a space
8035 
8036~rabsef yep - that's it
8037 
8038        we know that ?: has 2 children
8039 
8040        in the case of space, we would get a true value, right? So, we
8041        would then proceed to the first child
8042 
8043        line 11
8044 
8045~somlus if it is then begin another recursion
8046 
8047~rabsef right
8048 
8049~somlus oh no.
8050 
8051~rabsef exactly
8052 
8053        but what if we pass abc?
8054 
8055        no that's right - if we passed "space" our old value, which
8056        began with a space, we would be sent to line 11
8057 
8058~somlus ok...
8059 
8060~rabsef but if we do not - if we pass our old value "abc" which does not
8061        begin with a space we would not reach line 11, instead we would
8062        be sent to 12
8063 
8064        Here you might begin to see how we can have multiple bucs
8065 
8066~somlus I do not understand passing abc... you mean use as user input?
8067 
8068        or what is 'pass' here?
8069 
8070~rabsef I only mean inputting either " def" or "abc" as our old values
8071        were called (space and abc respectively)
8072 
8073        as the input to this program
8074 
8075        so if the input to this program were " def" we would do what
8076        happens at line 11
8077 
8078~somlus got it
8079 
8080~rabsef if our input to this program were "abc" however, we would not
8081        hit line 11
8082 
8083        and, as we were saying, if we don't hit line 11, we can see more
8084        clearly why we might be allowed to have multiple $s
8085 
8086        so long as we only allow for one $ to happen, /different/ things
8087        can happen in each recursion, based on those conditions that
8088        make the various $ mutually exclusive
8089 
8090        does that follow?
8091 
8092~somlus ah you have to continue the recursion but under different
8093        circumstances, so each needs a buc
8094 
8095~rabsef Yes - and lets think abck to the intent of this program
8096 
8097        "abc" -> "AbC"
8098 
8099        we are doing /different/ things to each letter in the tape
8100 
8101        sometimes upper, sometimes lower casing it
8102 
8103~somlus yes
8104 
8105~rabsef How often do we change this? Once every /other/ letter (or
8106        technically, we switch behavior each letter in the row)
8107 
8108        aha - perhaps we understand alternator now
8109 
8110        alternator helps us decide which to do
8111 
8112        let's look @ line 12
8113 
8114~somlus ok
8115 
8116~rabsef I'm not sure if we've discussed modular arithmatic yet
8117 
8118        basically mod x y gives us the remainder of x/y
8119 
8120~somlus did not discuss but I did read that.
8121 
8122~rabsef So (mod 1 2) will (sort of bizarrely because it's not really a
8123        remainder, but just take it as granted unless you want to talk
8124        about modular arithmatic) 1
8125 
8126        (mod 2 2) = 0
8127 
8128        (mod 3 2) = 1
8129 
8130        (mod 4 2) = 0
8131 
8132        noticing a pattern, we have a clear method of alternating
8133        behaviors
8134 
8135        0 1 0 1 0 1
8136 
8137        what does line 12 do then
8138 
8139~somlus interesting.
8140 
8141        It just flops back and forth between 0 and 1
8142 
8143        it's a toggle
8144 
8145~rabsef so long as alternator is increasing
8146 
8147~somlus on each recursion
8148 
8149~rabsef yes
8150 
8151        bingo bango
8152 
8153        so let's look at what happens if mod alternator 2 is 1
8154 
8155        that happens on line 13, as you have said
8156 
8157        let's break this into 3 parts - each separated by commas
8158 
8159        and first discuss the 1st and 3rd term
8160 
8161        what does the 1st do?
8162 
8163~somlus cool, this is toggle instruction to go to 13 then 14 alternately
8164        as long as it keeps going through that recursion space
8165 
8166~rabsef yep - and if we want to jump slightly ahead, condition 10 and
8167        result 11 handle the exception to the rule, where the value at
8168        the head of incomingtape is a space
8169 
8170~somlus hmm this is getting a bit complex, will take me a bit.
8171 
8172~rabsef we can't really capitalize spaces or lower case them so we had
8173        to build in an exception
8174 
8175~somlus makes senes
8176 
8177        sense
8178 
8179~rabsef So back to the thought -> the first term of 12 says "make
8180        incoming tape just the tale of tape as it exists coming into the
8181        rcursion"
8182 
8183        recall our pinned thought - we must lead to a termination point
8184        or we will build an infinite loop
8185 
8186        our termination point is when incoming string is null
8187 
8188        a tape is a list of cords, a list is always terminated with a
8189        null value
8190 
8191        combining those three things, we will always, eventually lead to
8192        a null list by pruning incoming tape to it's tail, no matter the
8193        length
8194 
8195        does that make sense?
8196 
8197~somlus overall yes, but I do not understand this line yet => the first
8198        term of 12 says "make incoming tape just the tale of tape as it
8199        exists coming into the rcursion"
8200 
8201        12 is a logic test
8202 
8203~rabsef $(<incomingstring t.incomingstring>,blah,blah)
8204 
8205~somlus oh I have that as 13
8206 
8207~rabsef that in <> is what i mean
8208 
8209~somlus cool ok.
8210 
8211~rabsef oh my bad dude
8212 
8213        it's 13
8214 
8215        I got confused
8216 
8217~somlus groovy, I figured.
8218 
8219~rabsef looking @ the third term of /13/ we see something else fairly
8220        easy to understand
8221 
8222        we know that alternator needs to increase to give us, well,
8223        alternation
8224 
8225        and here we increase alternator by 1
8226 
8227~somlus ah so incoming and outgoing are changing throughout the
8228        generator, little by little and both are sort of approximating
8229        the endpoint with each recursion
8230 
8231        yes that is cool and easy.
8232 
8233~rabsef precisely. This leads us to flop, weld and cass as we learned
8234        them earlier
8235 
8236~somlus and this is long format as compared to tall format. Would be
8237        interesting in education to show both formats.
8238 
8239~rabsef to give away the magic trick, we are going to pluck a value from
8240        the top layer of our incoming list (tape), do something with it
8241        based on the value of alternator at that instant, and then stick
8242        it to our outgoing tape
8243 
8244~somlus ok so here is another important point, which we sort of already
8245        stated, but not only should we state what we are going to do as
8246        part of an exercise, but what individual building blocks we will
8247        need.
8248 
8249~rabsef in this way, we can slowly prune incomingstring and slowly build
8250        outgoingstring
8251 
8252        you're right -we should do experiments and several of them with
8253        each standard library operation before hand
8254 
8255        I assume too much
8256 
8257~somlus So here, we need: recursion (including multiple endpoints that
8258        still trigger a recursion), an understanding of what a tape is
8259        and how to piece it apart character by character, something that
8260        alternates 0 to 1 over and over again
8261 
8262        yeah, this is great though. let's switch over to dm ok?
8263 
8264~rabsef we could build a simple naked gen that creates a list of
8265        alternator 0 1 0 1 and terms at a value of alternator proper (4
8266        cycles or whateveR)
8267 
8268        yah let's do it
8269 
8270        we can finish this next week or later
8271 
8272~somlus agreed
8273 
8274~risruc ~somlus, can you still read my DM?
8275 
8276~somlus yes sir
8277 
8278        I just answered it now. I'm surprised I missed that as I had
8279        been checking chat yesterday and today, feeling like I was the
8280        only one in this ghost town
8281 
8282~risruc no worried! I had the same feeling yesterday, compounded by my
8283        DM, I was afraid I wasn't properly sending or receiving. all
8284        good!
8285 
8286~rabsef Hey everyone - This week we're going to go through the SpOnGeBoB
8287        generator again as well as the Ackermann function available on
8288        the tutorials. It would be swell if new people want to join -
8289        I'm trying to improve my descriptive processes so happy to have
8290        you
8291 
8292~somlus Time for the hoontards session is 9am Sunday PST
8293 
8294        usually runs about 1.5 hours. It's technically 1 hour but
8295        usually it's rolling and fun and we don't stop until 1.5-2h
8296 
8297        hey ~rabsef maybe post that in /party and /nopub
8298 
8299~rabsef Yah for sure - do you happen to still have the pastebin link?
8300 
8301        also do you think the Tlon people will get pissed at all about
8302        it?
8303 
8304        I can regen the pastebin link - just trying to cut corners lol
8305 
8306~somlus I'm sure I can find it - which was the pastebin link the
8307        spongebob one?
8308 
8309~rabsef yes sir
8310 
8311~somlus https://pastebin.com/vK1DTLAS
8312 
8313~rabsef bingo bango
8314 
8315        thank you
8316 
8317~somlus I can't imagine they would be angry - they are super
8318        non-territorial. Oh you mean about posting in /nopub? Honestly I
8319        think they think this hoontards effort is sort of 'cute' and I
8320        do think they know that this level of learning has value, but is
8321        not the best place for their efforts at this time.
8322 
8323~rabsef That seems like a fair read to me
8324 
8325        I'm doin it - one sec while I write up an announcement
8326 
8327~rovnys i think you could simplify that code a bit by just toggling a
8328        bit instead of incrementing and calling +mod each time
8329 
8330        toggling a boolean flag, I mean
8331 
8332~rabsef that's a good point - I am not comfortable w/ loobeans yet but I
8333        shoulda done that
8334 
8335        I mean there's nothing difficult about loobeans I'm just not
8336        used to processing them as something to do
8337 
8338        ~rovnys - I didn't ask my question very well in ~dalten but
8339        I have an example now. Suppose ~dalten is "Steam" and it's
8340        owners/managers are video game producers. Could ~dalten act as a
8341        distributor of the producer's product. Like REI started out as a
8342        coop to do just that
8343 
8344        That's what I was trying to get at - assuming that the
8345        application qua game continued to 'live' on the underlying
8346        producer's machine could ~dalten provide a face for getting to
8347        that content
8348 
8349        maybe that's just an inefficient use of two machines; I'll take
8350        that answer too - just thinking aloud here.
8351 
8352~somlus are you able to get into the dalten coop? I can't get in
8353 
8354~rabsef Yes - ~somlus, next time we have a meeting I'll make sure
8355        you're apprised through our DM. Sorry I didn't think about
8356        the difficulties you were xpressing in U-H re that and I was
8357        on a work call up until 10 min /into/ the meeting today,
8358        unfortunately, so I missed some of it as well
8359 
8360        goddamned work getting in the way of my moonlighting
8361 
8362~rovnys ~rabsef I'm not sure, maybe
8363 
8364        technically, I don't see why not
8365 
8366        but we don't have any drm support in urbit and that might be
8367        hard to do
8368 
8369        not sure if that matters for your game distribution model
8370 
8371~rabsef follow up: do you think that's a dumb idea because why wouldn't
8372        the star just take the service and run from there
8373 
8374        eh - I hadn't thought that far into it, I was just trying to
8375        model the opposite of what you'd said in my brain a little re:
8376        ships running mirrors of dalten's events to ensure consensus /
8377        identify misbehavior
8378 
8379~rovnys oh, well I do think a star could be a reasonable way to
8380        arbitrate disputes
8381 
8382        and also as a sort of poor man's blockchain
8383 
8384        if you want lower latency and less hassle, but you still want
8385        a neutral third-party node to store a sequence of hashes or
8386        something
8387 
8388        you could use a star for that
8389 
8390        that would be a way for a star to ensure consensus and help
8391        prevent misbehavior among other nodes
8392 
8393        those nodes also wouldn't strictly need to be the star's
8394        sponsors, especially if you want someone really neutral, like a
8395        notary
8396 
8397~rabsef As a mild aside, I tried to sell my law firm on blockchain
8398        stamping wills and estate documents but they didn't quite get it
8399        and also it's not technically self-certifying evidence in all
8400        states yet (blockchain records, that is). I would be super into
8401        urbit moving to try to fill that space some day.
8402 
8403        as you say - like a notary
8404 
8405~rovnys oh that's interesting
8406 
8407~rabsef The two parts they failed to recognize is 1) it will be shortly
8408        self-certifying evidence in all states, and already is in some
8409        and 2) younger people would think it was dope and eventually
8410        young people who are rich are going to be our clients in that
8411        space (as they get older)
8412 
8413~rovnys it could be an interesting service to provide as a law firm,
8414        even if it doesn't yet mean anything officially
8415 
8416        yeah
8417 
8418~rabsef I think NH and MA already have that on the books re: 1
8419 
8420~rovnys i didn't know that
8421 
8422        seems reasonable
8423 
8424~rabsef http://blogs.asucollegeoflaw.com/lsi/files/2018/11/Gary-Marchant
8425        -Blockchain-Admissibility.pdf
8426 
8427        it's vermont and some others - to correct myself
8428 
8429        ^above covers some controversy over that finding but I don't
8430        think there's much avoiding it
8431 
8432        The software I wanted to use was called AmiStamp
8433 
8434        they also have a sick service - this one is real interesting
8435 
8436        So when you do deals (M&A stuff), you have data rooms for due
8437        diligence
8438 
8439        the problem is you want to be able to turn this off for bidders
8440        as they drop out right? But - obviously - they could just keep
8441        copies of the documents which gives them intelligence you don't
8442        want (obviously there's no real way around this w/ photographic
8443        memories etc, but better to not give them the chance)
8444 
8445        The AmiShare product by AmiStamp allowed you to give a drop-box
8446        like experience to all users but every file was wrapped in
8447        a proprietary filetype that would only unlock based on the
8448        blockchain and you could build in time-outs on it so like "you
8449        can read my documents for 6 months and no more" or "I've decided
8450        you're out of the running so cut all access to that user
8451 
8452        I think there's a kernel of awesome there
8453 
8454~rovnys huh, that is interesting
8455 
8456        I wonder if you could notarize encrypted documents using
8457        zero-knowledge proofs
8458 
8459        sorry, vague thought
8460 
8461~tocref Can't wait until my life simmers down sometime this weekend and
8462        I can finally submit the hoon 201 week 1 assignment lmao
8463 
8464~rabsef Hoon 201 is running?! I thought it was suspended
8465 
8466~tocref It's been a /while/, my friend
8467 
8468~rabsef Gonna have to google 0 knowledge proof now :)
8469 
8470        ah - well then I don't feel toooo bad but also consider me jelly
8471 
8472~ronreg at the last sf urbit meetup harrison at tlon said that he was
8473        working on what sounded like some major redesigns to hoon
8474 
8475        which sounded exciting, certainly, but also makes me a bit
8476        trepedatious about spending time on hoon right this second
8477 
8478~rabsef eh - if you get it once you can get it again
8479 
8480        I need to come up w/ a new project in hoon and start learning
8481        again. I'm sort of at a lull and the best way for me to learn is
8482        w/ an active goal
8483 
8484        I think I need to learn JS too so I can make it landscape active
8485        - I'd like to do that at least
8486 
8487~rovnys ~ronreg those changes won't land for a long time, I wouldn't
8488        worry about it
8489 
8490        and if they make hoon harder to use, we won't accept them :)
8491 
8492~somlus Am going to update the binary for ~somlus-savlev and ~nibfeb
8493        now, will be offline just a bit
8494 
8495        I am back, successful
8496 
8497        I was thinking the same thing ~ronreg after learning there was
8498        'a lot to be fixed' in hoon. discourages me from putting too
8499        much into building something big at this time. However, I still
8500        have lots to learn so maybe that's fine
8501 
8502~navpun Interesting, is there any overview of what needs major changing?
8503        wouldn't a major hoon overhaul involve changing some more basic
8504        stuff?
8505 
8506~somlus it was a very casual conversation. I took it to be a bit of a
8507        "curtis is gone and he left hoon with some stuff I have to fix"
8508        conversation. I'm not sure it was the same person referenced
8509        above so I don't want to incriminate anyone. we did not get into
8510        any details at all.
8511 
8512~rabsef as long as they keep hoon weird I'm ok w/ having to relearn.
8513        Relearning gives you an opportunity to re-consider the things
8514        you learned the first time and how you understand them
8515 
8516        Also ~rovnys was suggesting that maybe learning Nock would be
8517        helpful to understanding more advanced pieces of hoon, and if
8518        Nock isn't changing, then maybe that would be a good place to
8519        spend time
8520 
8521        if hoon expressions change, they still can only mean pieces of
8522        Nock
8523 
8524~somlus Good point ~rabsef
8525 
8526~risruc ~rabsef, did you see a while bag, ~taglux-nidsep shared his
8527        hand-written notes from writing some non-trivial program in raw
8528        nock?
8529 
8530        https://twitter.com/JRstract/status/1202800817767698432
8531 
8532~rabsef I did not see that - I need to take a look - do you happen to
8533        have that content?
8534 
8535        nice man
8536 
8537        thank you
8538 
8539~risruc the next message says he had several dozen pages like that
8540 
8541        seems like ~somlus is a bit slow this morning
8542 
8543        unless it's me.
8544 
8545~rabsef test
8546 
8547        no change in brave
8548 
8549        how very odd
8550 
8551        And I /just/ installed brave
8552 
8553~somlus yeah i am hanging again with messages not going out, browser not
8554        refreshing.
8555 
8556        Do you all have any idea when the next network breach will be?
8557        if 1 month or so I will just breach this weekend and see if that
8558        fixes it all. If soon, I'll wait for that.
8559 
8560~risruc yeah, interesting ~rabsef.
8561 
8562        i wonder how that's possible.
8563 
8564        i wonder how that's possible.
8565 
8566~rabsef ~somlus, given that they just rolled out 10.2 I would imagine
8567        some time. It would be surprising to me if they had a breaching
8568        update immediately following a non-breacher. I would assume they
8569        got a lot of their recent work packaged in those two OTA and
8570        10.2 updates.
8571 
8572~somlus yeah good point
8573 
8574        how much would it adversely affect y'all if we lost the
8575        hoontards chat history? ~rabsef I actually copied most of our
8576        sessions, FYI. I might breach and I'm not aware of chat history
8577        saving measures...
8578 
8579~risruc i think that we might keep our chat histories if you give the
8580        channels the same names
8581 
8582        but not positive about htat
8583 
8584~somlus notebooks are in home/web (on a mounted folder) as publish and
8585        its subfolders (notebooks)
8586 
8587        but I can't find anything that looks obviously like the chat
8588        folders to load up and commit
8589 
8590~risruc test
8591 
8592~rabsef Going way out on a limb, I suspect chats are in the event log
8593        and therefore do not have object permanence in their own file
8594 
8595~somlus could be, though I would be surprised they were entangled. No
8596        wonder there are long load times if that is true...
8597 
8598        aha it's on chat-store!
8599 
8600~rabsef dotpoast
8601 
8602        second dotpost
8603 
8604        o wow that worked
8605 
8606        I would not have expected that to work
8607 
8608        ~risruc is a genius
8609 
8610~risruc I had a hacked index.js for mobile use before the official
8611        mobile version came out
8612 
8613        that's how I knew where to go :)
8614 
8615~rabsef lol amazing
8616 
8617~somlus risruc does dazzle quite often.
8618 
8619~risruc i try to keep a good balance of dazzle and fizzle.
8620 
8621~tocref If anyone has JS questions feel free to DM me on here; I
8622        have experience with javascript as well as react.js (which
8623        urbit's web interface uses, and involves learning concepts and
8624        conventions that regular JS doesn't think about much)
8625 
8626~rabsef So - question - should I want to build a tile for urbit
8627        landscape I should learn the react framework?
8628 
8629~tocref Yes –– though apparently they built tools to help
8630        https://urbit.org/docs/tutorials/hoon/landscape-tile/
8631 
8632~somlus interesting. thanks tocref! you are certainly welcome any time
8633 
8634~tocref React also requires a good bit HTML/CSS knowledge to use, but
8635        you can learn all the useful HTML in a couple hours and CSS is
8636        something you never truly "learn"
8637 
8638~risruc yes, appareciated!
8639 
8640~tocref Though Javascript and React are easier to work through when
8641        something goes wrong because the userbase is large enough that
8642        if you have a problem someone else has already come across and
8643        solved it most likely
8644 
8645~rabsef nice. and thank you, like the others said. I need to get to that
8646        tutorial. I'm going to complete the numbered ones in the next 2
8647        weeks and get started on the advanced stuff
8648 
8649        I have reinvigorated my desire
8650 
8651~tocref I also believe that urbit uses a modified version of tachyons
8652        for it's CSS framework
8653 
8654        https://tachyons.io/
8655 
8656        I actually fixed a couple UI issues in the interface and it made
8657        me really like tachyons
8658 
8659~risruc that's great ~tocref
8660 
8661        what's your level of hoon expertise?
8662 
8663~tocref I finished 101, though I had a pretty ugly submission for the
8664        final assignment (it worked though)
8665 
8666        Hoon is pretty cryptic to me but I've wrapped my brain around
8667        most of the basic concepts
8668 
8669~somlus very cool. what drew you to hoontards other that good
8670        samaritanship?
8671 
8672~tocref I went through the list of landscape cities and joined a few of
8673        them
8674 
8675~rabsef ~tocref we should compare finals - mine was p ugly too :)
8676 
8677~tocref Also not really sure I'd consider it "good samaritanship" though
8678        I'm like, barely above the level most people here seem to be
8679 
8680~somlus I'm hoping to do hoon 101 for real this time. l
8681 
8682        I am hoping to do hoon 101 for real this time. I'm a non coder
8683        so I fell off last time out of confusion but risruc has rehabbed
8684        me significantly
8685 
8686~risruc it's the part you were born to play!
8687 
8688~tocref I feel like being a coder might make you bad at hoon but in a
8689        different way than being a non-coder
8690 
8691~somlus I have heard that a few times.
8692 
8693~tocref https://pastebin.com/XzsjT24N
8694 
8695        This was my final assignment for 101; There's two versions, one
8696        doesn't work
8697 
8698~rabsef https://pastebin.com/TL47Uunv
8699 
8700~tocref One like barely works and ocassionally takes like 10 full
8701        minutes to arrive at a result lmao
8702 
8703        Yours is way better than mine; I think they didn't introduce
8704        typing at that point in 101 when I did it
8705 
8706        Custom types I mean
8707 
8708~rabsef o u did something very interesting w/ the cards tho
8709 
8710~somlus I sometimes worry about Urbit's computing power. our level of
8711        programs should not be doing that
8712 
8713~rabsef Also I had help learning how to do custom types so
8714 
8715        doing wat ~somlus?
8716 
8717~tocref They all had us just make lists but the emoji + number was a
8718        really elegant solution that someone else in my cohort did
8719        successfully ~rabsef
8720 
8721~somlus I'm bowing out haven't done it yet. also turning in soon so
8722        cheers folks. big surgery day tomorrow
8723 
8724~tocref If you guys haven't done the ackermann function yet, please do
8725        not run it with numbers larger than 2
8726 
8727        It's designed to be computationally absurd and it will take
8728        literal hours/days/weeks with large numbers
8729 
8730~rabsef oh yah I didn't think about it but emoji I guess are UTF8?
8731 
8732~tocref The ones I used are technically pre-emoji UTF8 characters; but
8733        yes emoji are UTF8
8734 
8735        The real crazy use of utf8 is in swift; which has full utf-8
8736        support –– so you can actually use emoji for function names
8737        (which you shouldn't actually do)
8738 
8739        Though being able to set 3.1458 to π instead of the word pi is
8740        kinda cool
8741 
8742        I wanna get myself to the level where I can make things in
8743        landscape that interact with a hoon backend; because from
8744        reading the docs it seems to be a deeper integration than just
8745        async data calls
8746 
8747~rabsef like what would that look like?
8748 
8749        like a for instance of what you want to do
8750 
8751~risruc ~ponmep-litsem had built a hoon web editor, in the pre-landscape
8752        days.
8753 
8754        "plato"
8755 
8756        but he used elm, rather than react.
8757 
8758        that was so long ago, i'm guessing his elm-bindings don't work
8759        anymore
8760 
8761        but it was pretty impressive work.
8762 
8763~rabsef it's humbling to know so little :)
8764 
8765~risruc last time i'd checked this image worked... but doesn't seem to
8766        now.
8767 
8768        https://twitter.com/urbit/status/977619558117752834
8769 
8770        any way, it's still on his github:
8771 
8772        https://github.com/mikolajpp/plato
8773 
8774        https://github.com/mikolajpp/plato/blob/master/elm/plato-peek.gi
8775        f?raw=true
8776 
8777        there it is ^
8778 
8779        https://i.imgur.com/lzNnJJL.mp4
8780 
8781~hatdyl Urbit isn't exactly fast but it's not ludicrously slow. I
8782        wrote a procedural planet shader in shadertoy and then ported
8783        it to Hoon and it's slow but not completely impractical.
8784        https://gist.github.com/mattnewport/f3eaa71c3f71cafe83da541cf3c0
8785        52ae
8786 
8787~rovnys whoa, nice!
8788 
8789~hatdyl We're talking 1000s of times slower than a GPU which is 'only'
8790        like the difference between a 90s PC and a modern GPU
8791 
8792~rovnys I hadn't seen this before
8793 
8794~risruc that is awesome ~hatdyl
8795 
8796~hatdyl You could still do a lot with a 90s PC
8797 
8798~risruc can you quantify that at all?
8799 
8800~rabsef that's amazing ~hatdyl
8801 
8802~rovnys that's some pretty good-looking hoon, ~hadtyl
8803 
8804~somlus Is there hope that Hoon will every be computationally fast, or
8805        is that something we just give up and focus on "slow computing,
8806        but communities can be created"? Seems like a very core issue
8807        there to set long term expectations. BTW very cool planet and
8808        love the comment by Phillip and your reply on GH
8809 
8810~rapfyr hoon can be much faster than it is currently
8811 
8812        it's just not something that is being actively focused on.
8813 
8814        i mean other than jaque
8815 
8816~somlus got it, that is good to hear at least!
8817 
8818~rapfyr the previous version of jaque, iirc, was much faster than vere
8819        without any actual optimization work
8820 
8821~rovnys we'll probably never be full-speed on the processor, but if
8822        someone dumped the same amount of money into making hoon fast
8823        that people have dumped into making javascript fast, hoon will
8824        probably be quite a bit faster than javascript
8825 
8826        and javascript is actually not that slow for most things
8827 
8828~rapfyr i mean maybe if someone makes a nock chip
8829 
8830        zoom zoom.
8831 
8832~rovnys that is a possibility
8833 
8834~risruc i know a guy.
8835 
8836~somlus what a newsworthy item that would be...
8837 
8838        you do?
8839 
8840~rapfyr hey... hey you... you wanna... buy a nock chip?
8841 
8842~somlus oh, a joke. I know nothing about chip making
8843 
8844~risruc we're all just waiting for the loop to run enough times that the
8845        2025 time traveler has a nock chip to bring back to 1990 curtis.
8846 
8847        you weren't my guy, ~somlus.
8848 
8849        you're not my only guy you know ;)
8850 
8851~somlus I'm the bull in a china shop of chip making. First of all, I
8852        begin with Doritos
8853 
8854~rapfyr mmmmm nock flavored Doritos. my favorite.
8855 
8856~risruc how would you describe the flavor?
8857 
8858        [[[2 [3 [1 2]]...
8859 
8860~somlus it's actually a Doritos nock off. flavor has yet to fully
8861        develop
8862 
8863~rapfyr ...
8864 
8865        i'm dead
8866 
8867~hatdyl Because hoon is pure and side effects are pretty well defined
8868        I'm somewhat optimistic that purely computational stuff
8869        should be able to be sped up pretty dramatically. I don't see
8870        any reason in theory why it wouldn't be possible to offload
8871        something like this to the GPU for all the heavy lifting since
8872        it's side effect free until you write the output somewhere. I'm
8873        not an expert on this stuff though.
8874 
8875~tocref ~rabsef a data call is "asking another source for data" and
8876        asynchronous (async) means "do this on a seperate thread"
8877 
8878        in this context it means a frontend interface querying a
8879        backend via the internet –– think like a social media website
8880        periodically contacting the server for new notifications
8881 
8882~ronreg that procedural planet generator is super nifty
8883 
8884~somlus Getting ready for hoontards session to begin at 9am (35mins)
8885 
8886~tomdet somlus here, speaking as moon
8887 
8888~rabsef Hey ~somlus
8889 
8890        Updated SpongeBob to be by loobean instead of mod alternator ->
8891        https://pastebin.com/1TsUpnih
8892 
8893        https://pastebin.com/1TsUpnih
8894 
8895        Ackermann Function:
8896 
8897        https://urbit.org/docs/tutorials/hoon/ackermann/
8898 
8899        Anyone else joining us today?
8900 
8901~tomdet ready rabsef, sir, when you are
8902 
8903~rabsef Let's hit it
8904 
8905        So - last time we were working through the SpongeBob script
8906        above - and I think we got most of the way through
8907 
8908        I've made a small change to use loobeans instead of an
8909        incremental counter w/ mod operations
8910 
8911        you'll see that on lines 5 and 12-14
8912 
8913~tomdet interesting, ok. (also fyi because of other projects I come
8914        quite unread with regards to today, so we are winging it a bit
8915        on my end, apologies)
8916 
8917        ok focusing on thos lines
8918 
8919~rabsef Don't worry about it. So basically the biggest change is just
8920        that it's very easy to read an @f (loobean) as either & (true)
8921        or | (false) and, consequently, make it go the other direction
8922 
8923        (false or true, respectively)
8924 
8925~tomdet one moment, comparing to the old one now
8926 
8927        what does & do? change the alternator?
8928 
8929        oh no,
8930 
8931~rabsef & is a potential value for the alternator, as | is (false)
8932 
8933~tomdet it's true obviously and you're comparing alternator to & in 12
8934 
8935~rabsef yep
8936 
8937        line 12 we arbitrarily test for & (true) - it would do equally
8938        well to test for false and reverse the ordination of lines 13
8939        and 14
8940 
8941        Does that make sense?
8942 
8943~tomdet ah, then you just reset alternator to the other one right after
8944        you do the work in 13 or 14
8945 
8946        easy, peasy.
8947 
8948~rabsef Yes - so it seems like this whole generator is pretty much
8949        within your grasp now, yah?
8950 
8951~tomdet you set alternator as & and | yourself? or do they have inherent
8952        (immutable) meaning?
8953 
8954~rabsef They are immutable meanings and are just the way that Hoon does
8955        boolean values
8956 
8957~tomdet curious as to why you did not choose 0 and 1 or y and n?
8958 
8959        oh got it. So you can set a variable to be "true" or "false"
8960        this way. nice!
8961 
8962~rabsef that's right - and it just uses that notation 'because'
8963 
8964        or at least I don't have a better reason as to why
8965 
8966~tomdet cool
8967 
8968        great. this is easier.
8969 
8970~rabsef lil bit - yes. Thanks to whomever suggested that - I forget who
8971 
8972~tomdet ok great
8973 
8974~rabsef The big beat takeaway here is we have three different cases of
8975        buc where the resolution of the gate changes slightly based on
8976        conditions, right?
8977 
8978~tomdet yes, this is a very valuable learning tool.
8979 
8980~rabsef or, more simply, we can do different recursion strategies based
8981        on different conditions on each recursion
8982 
8983        let's then turn to the
8984 
8985        Ackermann function ^
8986 
8987        https://urbit.org/docs/tutorials/hoon/ackermann/
8988 
8989        four short lines but I would say somewhat more complex than what
8990        I did
8991 
8992~tomdet When describing this exercise, that should be stated in the
8993        'brainstorming' section. We will need to do something over and
8994        over (recursion), flip flop (recursion) and build something
8995        slowly as we tear something apart slowly (recursion)
8996 
8997        ok on to ackerman
8998 
8999~rabsef yes - that makes sense ^ noted
9000 
9001        ok - first let's read the bottom two paragraphs
9002 
9003~tomdet ok regarding ackerman, started doing this but did not get grasp
9004        the 'not primarily recursive' part FYI
9005 
9006        ok reading the bottom part
9007 
9008~rabsef Primatively Recursive is sort of helpfully defined by Wikipedia
9009        though it's not entirely necessary to grasp the definition to
9010        understand what's going on in the hoon there
9011 
9012~tomdet I cannot tell why this function would be useful
9013 
9014~rabsef In computability theory, a primitive recursive function is
9015        roughly speaking a function that can be computed by a computer
9016        program whose loops are all "for" loops (that is, an upper bound
9017        of the number of iterations of every loop can be determined
9018        before entering the loop).
9019 
9020~tomdet ah interesting.
9021 
9022~rabsef It's not useful to us, it's just an example of complex recursion
9023        for us
9024 
9025        so like my 'function' for spongebob (I think this is right) is
9026        primatively recursive because you know exactly how many loops
9027        based on length of tape incoming
9028 
9029~tomdet ok got it. I have entered that generator into my ship, committed
9030 
9031~rabsef ok - so let's look through the lines like usual
9032 
9033        and recall to our rules - line 1 - takes 2 arguments in a cell,
9034        both of which are @s
9035 
9036        ln 2 - a ?: operator which takes a test and two outcomes
9037 
9038~tomdet indeed
9039 
9040~rabsef Our first test - well that mirrors the first line of the
9041        penultimate paragraph
9042 
9043~tomdet easy, yes
9044 
9045~rabsef "If m is 0, return the increment of +n, else..."
9046 
9047        What's interesting here is that +n is going to be our output.
9048        How can I tell this?
9049 
9050        that is to say, this function will always end by outputting +(n)
9051 
9052~tomdet ah because there is no $ here
9053 
9054~rabsef and further because all other possible conditions are $ lines
9055 
9056        =IF(m=0, output, IF(condition, $, $))
9057 
9058~tomdet in terms of understanding this function, is m the number of
9059        overall recursions we will put n through (it counts down to zero
9060        and then the generator stops)?
9061 
9062~rabsef not exactly - there's a very good reason why not, too. We'll get
9063        to that
9064 
9065~tomdet ok
9066 
9067~rabsef So - first, if n = 0 we do what? recur with m-1 as m and 1 as n
9068 
9069~tomdet yes
9070 
9071~rabsef What's happening in the next line tho - line 4. Talk me through
9072        the things it's asking us to do
9073 
9074~tomdet ok.
9075 
9076        this is the case where neither m nor n equals zero
9077 
9078~rabsef that's true
9079 
9080~tomdet in this case first of all we have announced we will recur after
9081        doing the instructions. then we subtract one from m and make
9082        that the new value of m. and then somehow we are setting up a
9083        new recursion loop for n, but I can't see where the beginning of
9084        that loop is
9085 
9086~rabsef that's why m is not a counter of recursions
9087 
9088        that right there is the special part of ackermann
9089 
9090        let's focus on the n condition in line 4
9091 
9092~tomdet ok
9093 
9094        i'm with you.
9095 
9096~rabsef n will become the value of what? $ always points to |- above,
9097        so n will become the value of the ackermann function run with
9098        m=current n and n=current n - 1
9099 
9100        there is a mini-loop that will then play out the whole four
9101        lines again to determine the maxi-loop n value for recursion of
9102        the four lines
9103 
9104~tomdet oh wow. this is a little bit of turtles on the backs of other
9105        turtles.
9106 
9107~rabsef Eh - since |- doesn't exist it just points to the gate, but
9108        that's not that strange
9109 
9110~tomdet but not infinitely
9111 
9112~rabsef yes - it's turtles all the way down
9113 
9114        but it does take a lot of computer power to do this - if you put
9115        in relatively mild values of m and n, you're going to get back a
9116        relatively large number and it will take a while to work itself
9117        out
9118 
9119        I suggest you don't do it - but you can work this out on paper
9120        too - let's try with 1 1 on paper?
9121 
9122~tomdet ok
9123 
9124        I think I need to do that in order to understand this.
9125 
9126~rabsef It's going to take a bit of work to even write
9127 
9128~tomdet we get immediately to line 4
9129 
9130~rabsef yep
9131 
9132~tomdet m becomes zero
9133 
9134~rabsef so for our maxi recursion m is 0 for next loop (we know this
9135        means we will output soon!)
9136 
9137        but n is problematic - so somehow indicate a sidebar of a mini
9138        loop
9139 
9140        and what do we start the miniloop w/
9141 
9142~tomdet n is initially 1 and it is going to become the value of the
9143        ackermann solution (with m being zero now)
9144 
9145        or perhaps M doesn't change value until this mini loop is
9146        complete?
9147 
9148~rabsef not quite. For the mini loop we will start w/ m=n which is
9149        currently 1, and n = 0
9150 
9151        do you see why?
9152 
9153~risruc enjoying the walkthrough. great stuff. fyi, following along on
9154        and off. ??  
9155 
9156~tomdet because of dec n
9157 
9158~rabsef actually - risruc - am I right about that?
9159 
9160        I think it's calling n and (dec n) as m and n in the recursion
9161 
9162        second half line 4
9163 
9164        but now that I think about it it might be just saying keep m the
9165        same and recur w/ n as dec n
9166 
9167        which in our case is the same thing but it's sort of important
9168        that we think of it right
9169 
9170        It must be that second case
9171 
9172        so you're right somlus - m does not change, it retains it's
9173        value from the prior maxi-loop and then recurs with n = dec n
9174 
9175~tomdet I think it is keep m the same because buc is outside the
9176        parentheses
9177 
9178~rabsef otherwise it would need to be written n $(m n, n (dec n))
9179 
9180~tomdet this mini recursion happens in the environment before the $ line
9181        is complete
9182 
9183        yes
9184 
9185~risruc correct. in the final $, m doesn't change.
9186 
9187~rabsef yah - my bad - I was thinking of it wrong. either way, we can
9188        proceed w/ our mini recursion starting at the gate with m=1 n=0
9189 
9190        which takes us where
9191 
9192~risruc in the penultimate $, m becomes m-1 and n becoms the result of
9193        the given recursion with n=n-1
9194 
9195~tomdet so you have to "end the generator" in a mini sense by getting M
9196        to zero before you can complete line 4
9197 
9198~risruc part of the problem of the ackerman function as an illustration,
9199        in my opinion is that it's an abstract and silly function. not
9200        one that most people have a map of in natural language.
9201 
9202~tomdet true my head is spinning a bit
9203 
9204~risruc compared with factorial / fibonacci which are easy to describe.
9205 
9206~rabsef yep - and in that mini recursion w/ m=1, n=0, we go to line 3
9207 
9208~tomdet yes, I see that
9209 
9210~rabsef n does = 0, so we should mini recur with m=0, n=1
9211 
9212        which, 2nd (?) pass of the mini loop (I'm not actually writing
9213        it sorry) takes us to line 2
9214 
9215        and we know line 2 is an output line, thank god which means soon
9216        we will return to our maxi-loop
9217 
9218        thus, for our maxi loop, what is the value of n now?
9219 
9220~tomdet and so that n is finally n+1 which is 2 which is the result for
9221        that part, which means n is now 2
9222 
9223        for the maxi loop to begin again with m =0 and n=2
9224 
9225~rabsef and so
9226 
9227        ?
9228 
9229~tomdet it goes to line 2, which is true which produces the +(1) as 2 as
9230        the final output
9231 
9232~rabsef not quite
9233 
9234~tomdet dang
9235 
9236~rabsef for our maxi loop m=0, n=2
9237 
9238        what is our output on line 2 then
9239 
9240~tomdet 3 but let me think on why n=2 for the maxi loop
9241 
9242~rabsef let's return to our miniloop
9243 
9244        on the second pass what did we get as output?
9245 
9246~risruc ~somlus, have you looked at this, from wikipedia?
9247 
9248~tomdet oh yeah, we went into that loop with n=1 and so +(1) is too
9249        there
9250 
9251~risruc https://wikimedia.org/api/rest_v1/media/math/render/svg/1a15ea2f
9252        cf1977e497bccdf1916ae23edc412fff
9253 
9254        oops, that image won't embed.
9255 
9256~tomdet yes, risruc, but only briefly unfortunately. I sadly bogged
9257        myself down with other stuff and didn't really come prepared as
9258        I would have liked
9259 
9260        I have it open now though
9261 
9262~rabsef that's fine - you got to where we needed to be
9263 
9264        So - I would say try your gen w/ 1 1 and see if we get what we
9265        thought
9266 
9267        you could probably rewrite this generator to create a list of
9268        cells of [@ @] that showed you each pass output to coincide w/
9269        your maxi and miniloop recursion values
9270 
9271        which might be sort of interesting
9272 
9273~tomdet yeah actually it would
9274 
9275~rabsef you could even do it w/ a cord or term of %mini and %maxi [%mini
9276        1 0] [%maxi 0 2] and so on
9277 
9278~tomdet ah got it! had a problematic input format. it needs to match the
9279        line where the variables are defined.
9280 
9281~rabsef yes - ah and that's always true
9282 
9283        right? we have to give the system what it expects or we get a
9284        nest fail I think or a want has if a say generator
9285 
9286~tomdet cool it is working for now. got 29 for [3 2]
9287 
9288~rabsef want have* for say generator error
9289 
9290        yah nice - and we could work that out on paper but absolutely f
9291        that
9292 
9293        lol
9294 
9295        go too much bigger and it will bog your urbit down for a while
9296 
9297~tomdet haha!
9298 
9299~rabsef I tried something fairly big when I first wrote this and
9300        realized that the 'hard way'
9301 
9302~tomdet it did. had to interrupt [4 1]
9303 
9304~rabsef yah lol
9305 
9306        the poor computer has to go nest nest nest nest nest nest just
9307        like we do
9308 
9309        until it gets a value
9310 
9311~tomdet incidentally I am running planetppm at 1.024 and will see how
9312        that goes. it's at line 179 now
9313 
9314        cool, ok.
9315 
9316~rabsef is planetppm the planet shader?
9317 
9318~tomdet yes
9319 
9320~rabsef or ?
9321 
9322~tomdet I think I got this. I want to spend some time doing more
9323        handwritten versions and also go back to spongebob sarcastic
9324        text and make sure I get both versions.
9325 
9326~rabsef nice - do you want to do anything more today?
9327 
9328        we could look @ the caesar cipher
9329 
9330        that is going to take at least one week
9331 
9332~tomdet sure!
9333 
9334~rabsef https://urbit.org/docs/tutorials/hoon/caesar/
9335 
9336        let's just try to read through the code at the top
9337 
9338        new notation for us =<
9339 
9340~tomdet my browser froze, stand by
9341 
9342~risruc if we'll split it over weeks, might be useful for today to try
9343        and understand it at a higher level.
9344 
9345~rabsef 2
9346 
9347~rabsef =<(b [a=1 b=2 c=3])
9348 
9349~risruc break it up into black boxes
9350 
9351~rabsef Yes - I just want to get through lines 1-7 today
9352 
9353~risruc good call
9354 
9355~rabsef I guess we could include 9 lol 9's pretty easy to understand
9356 
9357~tomdet ok i have it now, beginning
9358 
9359~rabsef :)
9360 
9361        I guess we should begin at the beginning really
9362 
9363~tomdet good point since !: is actually new to me right now
9364 
9365~rabsef !: zapcol gives us a "stack trace"
9366 
9367        The product of p unless p crashes, in which case a stack trace
9368        is given.
9369 
9370        ^per the documentation
9371 
9372~tomdet got it
9373 
9374~rabsef basically "Do everything after me else give me a list of the
9375        lines you went to and where you failed"
9376 
9377        that's extremely helpful when ur debugging, incidentally
9378 
9379~tomdet so you could do that with every single generator
9380 
9381        got it
9382 
9383~rabsef yah for sure
9384 
9385~tomdet so the we have |= where we state that our final output which we
9386        call "msg" is a tape and is cast as @ud
9387 
9388~rabsef ok so line 2 we're familiar w/
9389 
9390        yep
9391 
9392~tomdet =<
9393 
9394~rabsef what would our input look like - arbitrary input value
9395 
9396        ["like this perhaps" 2] right?
9397 
9398        yes - now =< does something new as well
9399 
9400        2
9401 
9402        =<(b [a=1 b=2 c=3])
9403 
9404        that's an example of what it does that is new
9405 
9406        can you intuit from that what it does?
9407 
9408        or deduce I suppose would be a better word
9409 
9410~tomdet produce the value from a database
9411 
9412~rabsef sooort of. =<(p q) evaluates p in light of q
9413 
9414        where normally in hoon we would have to do q first
9415 
9416        so if you went into dojo right now and hit b
9417 
9418        it would be find.b error
9419 
9420        if you =b 2 then b it would give you 2
9421 
9422        =<(b [b=2]) will give you 2 in one line
9423 
9424        =< evaluates the first statement in light of the remaining
9425        information
9426 
9427~tomdet ok
9428 
9429~rabsef the /rationale/ for doing this is that you can have your main
9430        (output focused) work at the TOP of your program and have your
9431        other work at the bottom (everything that begins with ++ below)
9432 
9433        like a christmas tree w/ the star on top where the star is the
9434        important part, your branches can exist below it. Arguably your
9435        branches should be bigger than the star, and it's nice to look @
9436        w/ them at the bottom
9437 
9438~tomdet I understand the concept but not how =< achieves that (yet)
9439 
9440~rabsef does that track fairly well?
9441 
9442~tomdet oh
9443 
9444~rabsef that's just what =< does. Our lines 4-6 will be evaluated based
9445        on 7+
9446 
9447~tomdet what I need to grasp is where the endpoint of =< is. in your
9448        example you had an open and closed parenthesis
9449 
9450~rabsef Ok - great question, and I /think/ the answer is "after there is
9451        definitive output"
9452 
9453~tomdet oh
9454 
9455~rabsef ~risruc - weigh in if I'm wrong there
9456 
9457        let's try to find that point
9458 
9459        =. does something new too, to move to the next line
9460 
9461~tomdet ok i'll get back to =<
9462 
9463~rabsef =. 'mutates' the subject in some way and the rest of the
9464        function is evaluated in light of that mutated subject
9465 
9466        our subject in this case is the sample (and context but it isn't
9467        used)
9468 
9469        the sample is our input
9470 
9471~risruc that is correct, essentially.
9472 
9473~rabsef =. is doing something like $ here w/o recursion, but instead
9474        continuing to flow forward
9475 
9476~risruc although of course we don't need to see the output.
9477 
9478        think about the fact that each function (rune) takes a fixed
9479        number of children
9480 
9481~rabsef so what does =. look like it's doing
9482 
9483~risruc without evaluating anything, if we know that, we can see where
9484        the next hoon ends.
9485 
9486~rabsef that's a good point - so at termination of computing is where =<
9487        ends
9488 
9489~tomdet where would i find out what 'cass' refers to. is that standard
9490        library?
9491 
9492~risruc hoon meaning the rune after =< and its children
9493 
9494~rabsef yes sorry and let's give you a freebie there - cass is all lower
9495        case
9496 
9497        cuss is upper case
9498 
9499~tomdet ah got it.
9500 
9501~rabsef we used those in the SpongeBob - I should have mentioned that
9502 
9503~tomdet understood
9504 
9505~rabsef that's how we were getting the letters to flip case regardless
9506        of their input style
9507 
9508~tomdet got it I see that now
9509 
9510~rabsef ok so we take in a ["sample" 2]
9511 
9512~tomdet ["ucorng" "qyknjc"]
9513 
9514~rabsef and cass the "SAMPLE" into "sample" at line 4 using =. which
9515        'mutates the subject' for further computation. Technically, I
9516        suppose, the subject of msg="SAMPLE" still exists somewhere but
9517        it is not used hereafter
9518 
9519~tomdet got it.
9520 
9521~rabsef ok - and you already got output so that pretty much tells us
9522        what 5 and 6 do
9523 
9524~tomdet that's an easy way to convert to all lower case
9525 
9526~rabsef without worrying to omuch about what's in the ()() what is :-
9527        doing
9528 
9529        lmk if you want help but we've seen this rune before
9530 
9531        I think
9532 
9533~tomdet yes we have
9534 
9535~rabsef think of it w/o evaluating that in parentheses next to/below it
9536 
9537~tomdet construct a cell, and in this case two things will be there
9538 
9539~rabsef what is :- doing then
9540 
9541        yes - and those two things are going to be that in parentheses
9542 
9543~tomdet which we will define later, but will include a shift and unshift
9544        version of message key
9545 
9546~rabsef so [(shift msg key) (unshift msg key)]
9547 
9548~tomdet indeed
9549 
9550~rabsef yes exactly. and here we see why =< is used
9551 
9552        there are arms called shift and unshift just below this - in a
9553        core. msg and key are sample inputs
9554 
9555        so the output of =< will be the cell of "operate shift on
9556        sample" and "operate unshift on sample"
9557 
9558~tomdet ah
9559 
9560        yes
9561 
9562~rabsef look at the gate for shift - line uhh 10?
9563 
9564        what does it ask for
9565 
9566~tomdet an input which is assigned to message
9567 
9568~rabsef and a key right?
9569 
9570        it asks for a tape and an @ud
9571 
9572~tomdet indeed just like msg
9573 
9574~rabsef the notation (function arg arg2) passes the args to the function
9575 
9576        (shift msg key) passes (cass msg) and key from our initial input
9577        to shift
9578 
9579~tomdet ok
9580 
9581~rabsef does that make sense?
9582 
9583~tomdet just a moment, i think so
9584 
9585        when you say 'passes' here I think we need to be more specific.
9586 
9587        becomes? is equal to? does something to?
9588 
9589~rabsef gates require input |= [msg=tape key=@ud] on line 2 requires a
9590        cell of [tape @ud] as input
9591 
9592        when I say passes I mean calls the gate w/ the following
9593        arguments
9594 
9595        just as we would call +caesar [tape @ud]
9596 
9597~tomdet ok got it
9598 
9599~rabsef shift is a gate, see line 11
9600 
9601        we are +shift [tape @ud]
9602 
9603        only (shift tape @ud) serves the same purpose here
9604 
9605~tomdet got it
9606 
9607~rabsef and we can only call shift up here before it exists because =<
9608        says evaluate me in light of "q" which is the core at the bottom
9609        (lines 7+)
9610 
9611~tomdet that makes sense now.
9612 
9613        ok got it.
9614 
9615~rabsef let's look at line 8
9616 
9617~tomdet so coders like to have a neat summary of everything that will
9618        happen at the beginning before they get into the weeds.
9619 
9620        that is an important point for learning.
9621 
9622~rabsef I think that's the idea
9623 
9624        yes it is - good point
9625 
9626~tomdet A coder can look at line 1-6 and tell what is going to happen,
9627        and then they then know what to look for.
9628 
9629~rabsef it's not always true but it's a nice way of writing something
9630        simple like that
9631 
9632~tomdet as compared to the result of the entire effort being in line 42
9633        or something.
9634 
9635~rabsef yes and you can keep doing that in the core where shift calls
9636        operate, passing message key and the value of the arm encoder
9637        when passed key
9638 
9639~tomdet it is time efficient to do that. ok i'm getting this now.
9640 
9641        it's nice that it is set up that way. Basic as you know is not
9642 
9643~rabsef inside of a core, you can call anything that is part of the
9644        "subject". The subject of an arm of a core is the parent core
9645 
9646        that is to say shift can call unshift, operate, encoder,
9647        decoder, coder, zipper, and rott
9648 
9649        the core starts at |%
9650 
9651~tomdet got it. briefly define core please.
9652 
9653~rabsef yah.. about that
9654 
9655        lol
9656 
9657~tomdet is this where all of the "workhorses" are stored?
9658 
9659~rabsef It's very hard to define, for me. a Core is a broad definition,
9660        a gate is a type of core as is a door
9661 
9662        let me get their definition one sec
9663 
9664~tomdet ok
9665 
9666~rabsef A core is a cell of a battery and a payload. It is a Hoon
9667        construct similar in use to Objects in Object Oriented
9668        programming languages.
9669 
9670        remember that battery = [sample context] and payload is "what to
9671        do w/ it"
9672 
9673        in a simple gate |= a=@ud +(a), for instance, the battery is the
9674        sample of a and the payload is +(a)
9675 
9676~tomdet ok
9677 
9678~rabsef in this specific case our |% core is unlike a gate in that it
9679        doesn't take direct sample, but the context is the larger gate
9680        that took in msg and key
9681 
9682        (I think - ~risruc correct if I'm wrong therE)
9683 
9684        the context is something like the subject in which the thing
9685        occurs
9686 
9687        so the subject of our core is the 'program' that it's running
9688        in which includes the information coming in and given a face at
9689        line 2
9690 
9691~risruc it's useful enough to thing of a core as an object with one or
9692        more arm.
9693 
9694        a gate is a specific core with only one arm
9695 
9696~rabsef yah and a gate is one w/ one arm called?
9697 
9698        $
9699 
9700~risruc so it doesn't need a ++ arm declaration
9701 
9702~tomdet ok
9703 
9704        all right so |% defines a core for this cypher?
9705 
9706~rabsef yes - you could say it just starts a core period
9707 
9708~tomdet and then the items under ++ are called what?
9709 
9710~rabsef arms
9711 
9712~risruc ++ creates an arm of the core.
9713 
9714~posbus test (send ack if possible)
9715 
9716~risruc ack
9717 
9718~rabsef Yah - to clarify, the stuff under the ++ is /part/ of the
9719        declared arm
9720 
9721~tomdet cool
9722 
9723        hello pos
9724 
9725~posbus ty
9726 
9727~rabsef ok - I think we're probably done there for this week?
9728 
9729~posbus Hello everyone
9730 
9731~rabsef unless you have further questions
9732 
9733        ~posbus-bictud Hey bicdude
9734 
9735~tomdet I think so. I have a good amount to review
9736 
9737~risruc you can think of the arm ++ shift as you would a generator named
9738        shift.hoon
9739 
9740        you see from its gate that it requires a sample of [tape @ud]
9741 
9742~tomdet very helpful, and was wondering that. Thank you
9743 
9744~risruc and it works the same way with arms of a core.
9745 
9746        also, useful to review that "arm" simply means a wing with hoon
9747 
9748~rabsef the only important distinction between separate .hoon files in
9749        gen is that you can call the arms directly shift calls operate
9750        and so on
9751 
9752~risruc whereas "leg" means a wing with only data.
9753 
9754        remember that all data structures are binary trees.
9755 
9756~rabsef yah - that's a hard distinction imo. for instance, is alpha an
9757        arm or a leg?
9758 
9759~risruc each branch is a wing, either an arm(has hoon code) or a leg(has
9760        only data).
9761 
9762~tomdet ok I got all of that down, will study. Thanks guys!
9763 
9764~risruc good question, ~rabsef. by the definition i gave, we should say
9765        leg.
9766 
9767~rabsef but it's called as an arm I think maybe it's transiently an arm
9768        until evaluated and is then a leg because it has no evaluation
9769 
9770        ?
9771 
9772        lol
9773 
9774~tomdet This is nice because it tees me up for being ready for hoon 101
9775 
9776~rabsef there are a lot of dope caesar exercises
9777 
9778        I felt very accomplished when I figured some of them out
9779 
9780        I strongly recommend doing all of the exercises in the readings
9781 
9782        i suppose that's obvious
9783 
9784        ok I'm going to work on my conversion utility and then maybe a
9785        few more simple generators for the thing
9786 
9787~tomdet agreed, I really need to dive into that with gusto
9788 
9789        thank you for your time! makes me invigorated to dive in
9790        further.
9791 
9792~tomdet nice.
9793 
9794~polmex did anyone launche the Fake ship ? I can't. When I follow the
9795        aerology instructions video it appears this on my shell $ git
9796        clone https://github.com/urbit/urbit.git fatal: destination path
9797        'urbit' already exists and is not an empty directory.
9798 
9799        did anyone launche the Fake ship ? I can't. When I follow the
9800        aerology instructions video it appears this on my shell $ git
9801        clone https://github.com/urbit/urbit.git fatal: destination path
9802        'urbit' already exists and is not an empty directory.
9803 
9804        did anyone launche the Fake ship ? I can't. When I follow the
9805        aerology instructions video it appears this on my shell $ git
9806        clone https://github.com/urbit/urbit.git fatal: destination path
9807        'urbit' already exists and is not an empty directory.
9808 
9809        sorry for spam, just crashed lol #hoontard
9810 
9811~botter you already have something named urbit in that directory
9812 
9813~polmex I know, its my planet
9814 
9815~botter mk newdir
9816 
9817        then
9818 
9819        cd newdir
9820 
9821~polmex wherever?
9822 
9823~botter run your commands now and there'll be no conflict
9824 
9825        yeah. I'm assuming you're in the your home directory
9826 
9827        your default directory when you login
9828 
9829        this would be /home/[whatever your username is/newdir
9830 
9831~somlus thx for helping botter. welcome polmex
9832 
9833~botter er, the command is mkdir, not mk.
9834 
9835~polmex yep
9836 
9837        I fig it out
9838 
9839        I done it
9840 
9841        I know basic linux stuff but when it comes to affect my urbit
9842        thing I'm very cautious, I don't want to spoil my ship hehe
9843 
9844        thanks sir
9845 
9846~somlus that is concern well placed, polmex.
9847 
9848~polmex after cloning I can't either bootstrap
9849 
9850        when im in 'urbit' and type "./scripts/bootstrap" >> bash:
9851        ./scripts/bootstrap: No such file or directory
9852 
9853        :S
9854 
9855        guessing it could be because I need to install nix before?
9856 
9857~digmel Does it take a while for the notebook to show up? I was in last
9858        week, can't find my OG ship, idk how to run a fake one- so I
9859        deleted everything and started over- and here I am :) So- I
9860        ran the :publish &publish-action [%subscribe ~rapfyr-diglyt
9861        %hoon-school-assignments] command, and I got the error that
9862        everybody else did, but I also got >= which means successful
9863        right?
9864 
9865~risruc was your OG ship a planet or another comet?
9866 
9867~digmel well I am not entirely sure- I booted it up using my planet
9868        creds, but it gave me a different handle when I was in here- so
9869        I assume it was a comet- also during the boot sequence it said-
9870        if you want it to go faster get a planet- so... 53% confident in
9871        my answer: comet.
9872 
9873        I can't get back far enough in the chat to find my
9874        original original- i was also in with this handle:
9875        ~latnep-tacweb-sannev-lidtux--donmyn-mostug-noldev-marzod
9876        From my understanding I am creating a new comet each time-
9877        I do not know how to retrieve or relaunch an instance. Any
9878        recommendations?
9879 
9880~risruc was it 16 syllables like your current name?
9881 
9882        or 4 like mine
9883 
9884        4 = planet
9885 
9886        16, comet
9887 
9888        if you launched it successfully the first time, for subsequent
9889        launches, you don't need the -w or -k commands.
9890 
9891        those might result in mining a comet rather than using your
9892        plant
9893 
9894        in any event, since you've deleted everything, to get your
9895        planet back, you must log in to bridge and reset your network
9896        keys
9897 
9898        (breach continuity)
9899 
9900        under the admin section.
9901 
9902        then download your arvo keyfile
9903 
9904        and ./urbit -w sampel-palnet -k keyfile
9905 
9906        again, every time after that first boot, you will launch just
9907        with ./urbit sampel-palnet
9908 
9909~digmel OO! OK- I see, I didn't go back far enough to reset! Thanks.
9910        Before I go- if I wanted to come back using this comet- would I
9911        be able to? Not if I reset the key right?
9912 
9913~risruc yes
9914 
9915        just check the folder name it's using for data
9916 
9917        (its pier)
9918 
9919        and then run ./urbit comet-folder
9920 
9921        comets don't have keys except for their own names
9922 
9923        its name is a 128bit public key
9924 
9925        so you can't reset a comets network keys
9926 
9927        that's why they aren't meant to be permanent identities
9928 
9929        but as long as you don't delete the folder it's running from,
9930        you can use it forever
9931 
9932        or sink it somehow (screw up to the point you can't boot)
9933 
9934~digmel ok, thank you! I will try to be back as my planet.
9935 
9936~risruc planets you can sink as often as you like aaaand not lose your
9937        identity, since you can always reset your keys
9938 
9939        just ask our good friend ~rabsef :)
9940 
9941        I won't be here if you make it back. knocking out.
9942 
9943        good luck!
9944 
9945~digmel night!
9946 
9947        ty
9948 
9949~rabsef 22 rifts and counting. Probably need to do another one soon.
9950 
9951~lagrev needs a leaderboard
9952 
9953~rabsef I think in the .9 lineage I racked up 19 of those 22 and when
9954        they rebooted zod (or livpub or somebody) I displayed 19 times
9955        and several Tlon people were like wtf dude
9956 
9957~lagrev if you're not double digits do you even Mars
9958 
9959~rabsef go hard and do terrible things to your planet or go home. I
9960        learned it from you dad
9961 
9962~somlus comet digmel, did you make it back as your planet?
9963 
9964~polmex after cloning I can't either bootstrap when im in 'urbit' and
9965        type "./scripts/bootstrap" >> bash: ./scripts/bootstrap: No such
9966        file or directory
9967 
9968        any idea guys?
9969 
9970~somlus sorry if you answered this already but why are you going that
9971        route as compared to the curl and tar commands to get the files?
9972        I don't have much experience building binaries this way
9973 
9974~walden ~polmex-nocten: you want to follow the directions here instead:
9975 
9976        https://github.com/urbit/urbit/#development
9977 
9978        sorry about the double post, was having trouble reaching
9979        ~somlus-savlev
9980 
9981        i only posted once, fwiw
9982 
9983~somlus no worries, I only see one on my end. I have learned that a
9984        double post *can* occur when the sender is an a weak connection
9985        for some reason, but we have had somlus slight lag issues for a
9986        while now. it's much improved but occasionally still an issue.
9987 
9988~ronlyr .
9989 
9990~haclex i. am. in.
9991 
9992~somlus kaaaaaaaboom. welcome
9993 
9994~haclex i'm entirely useless in here of course, but it feels good to get
9995        in after being locked out all the same ;)
9996 
9997~somlus actually, I was thinking that you will be astounded with the
9998        extremely basic level on which we cover hoon. You might find
9999        it frustrating because you will likely grasp these concepts
10000        quickly. In any case you are, of course, more than welcome to
10001        participate any time you like.
10002 
10003~haclex there's presumably some sort of hoon tutorial on urbit.org
10004        somewhere in some form
10005 
10006        at least i've seen it referred to iirc. i figure i'll start with
10007        that at some point when my brain is functioning *and* i don't
10008        need it for doing more pressing things
10009 
10010~somlus I"m sure you found it by now but yes this is that section:
10011 
10012        https://urbit.org/docs/tutorials/hoon/setup/
10013 
10014~lagrev IMO there's a bit of a gap in the Hoon curriculum. There's a lot
10015        of time on basics (great!) and Gall development (great!) but
10016        there's not really an "advanced" Hoon or a "Hoon in practice"
10017        module.
10018 
10019        (The docs do cover this.)
10020 
10021        One of the consequences is that one often doesn't know when to
10022        dig for a library, use an existing pattern, or write something
10023        new.
10024 
10025~somlus yes lagrev, they are actually aware and we are going to reach
10026        out to see if they are open to ideas to help fix this! It's
10027        a fairly widely commented thing. There's a chance that some
10028        changes will come in time for the next hoon 101 session (not
10029        this one), but that's all pending. Sorry to be secretive-ish but
10030        don't want to set any expectations until we know if there is an
10031        interest in hearing creative solutions.
10032 
10033~lagrev awesome
10034 
10035~haclex i expect it to be sufficient for me to read an introductory
10036        text, map all of the functional concepts to whatever they
10037        renamed them to, and then read actual code. but i'm an edge
10038        case, so what you're doing here is great
10039 
10040        an architectural overview would be super helpful though, since
10041        trying to induce it from code is going to be a pain and a time
10042        sink i can't afford
10043 
10044~ronreg which is actually a neat thing to have a programming language
10045        std library, and of course hoon itself needs something like that
10046        to parse and compile hoon
10047 
10048        a few weeks ago I was trying to write a parser in hoon, and
10049        I was asking some basic question pertaining ot this, when
10050        someone pointed out that the std library has some basic parser
10051        combinator functionlaity already
10052 
10053~rabsef ~haclex could you point me to an architectural overview of
10054        something (anything) that you think is a good example of what
10055        you mean?
10056 
10057~hastuc https://urbit.org/docs/reference/vane-apis/gall/
10058 
10059        This is a pretty good hands on overview of the general structure
10060        of userspace applications
10061 
10062        (disclaimer: I wrote it)
10063 
10064        alternatively for a more high level view of how exactly urbit
10065        works the whitepaper is good
10066 
10067        https://media.urbit.org/whitepaper.pdf
10068 
10069~rabsef Thanks :)
10070 
10071~hastuc alternatively have a look through the urbit-dev google group
10072        it's a fantastic resource
10073 
10074~haclex ~rabsef an example of what i mean when i talk about a good
10075        architectural overview of a complex system is the linux
10076        netfilter docs rusty russell wrote around 2.4-2.6, complete with
10077        following the path of a packet through the linux kernel and
10078        tcp/ip stack
10079 
10080        i've been pointed to some urbit docs now, have not yet read them
10081        and don't know how good they are, but they're already better
10082        than what existed last time i asked since that was virtually
10083        nothing ;)
10084 
10085~rabsef I'll take a look @ that - thanks haclex
10086 
10087~haclex it's significantly outdated now in the era of ebpf, new-mm and
10088        so on
10089 
10090        it's significantly outdated now in the era of ebpf, new-mm and
10091        so on it's significantly outdated now in the era of ebpf, new-mm
10092        and so on it's significantly outdated now in the era of ebpf,
10093        new-mm and so on outdated now in the era of ebpf, new-mm and so
10094        on, but worth a read anyway
10095 
10096        wow, wasn't *so* good it was worth saying that many times
10097 
10098~lagrev haclex, do you need some percussive maintenance to recalibrate
10099        your vocoder?
10100 
10101~haclex either that, or for urbit to handle network hiccups better, or
10102        for my cable headend to finally figure out these 8 channels are
10103        noisy and give me new ones
10104 
10105        hmm, I didn't think I posted that to this channel
10106 
10107~somlus Question for ~rabsef or ~risruc or whoever:
10108 
10109        Is there a reasonably straightforward way to explain what the
10110        "expands to" components of the standard library runes means?
10111 
10112~risruc specific exampel?
10113 
10114~somlus one moment
10115 
10116        https://urbit.org/docs/reference/hoon-expressions/rune/bar/
10117 
10118        all runes have the same sections: syntax, semantics, expands to,
10119        discussion, example
10120 
10121        I am trying to grasp these different sections, though I know
10122        most of it is over my head. as long as i understood the intent
10123        of the expands to section I would be further along
10124 
10125~risruc "expands to" means "could be written with lower level runes as:"
10126 
10127~somlus ok thanks.
10128 
10129        perfect
10130 
10131~risruc not only "could be written" but "is interpretted as"
10132 
10133~somlus I do not yet know those lower level runes so there is the
10134        confusion
10135 
10136        got it - to build this command, we used these simpler tools so
10137        you don't have to every time...
10138 
10139~risruc it's like saying 5! expands to 5*4*3*2*1
10140 
10141        correct ^
10142 
10143~somlus ok great. that will actually help me one day understand the
10144        lower level commands because there are so many examples of using
10145        them of course.
10146 
10147~risruc protip: print out the cheat sheet
10148 
10149        or just have it open in a tab at all times
10150 
10151~somlus I have it but still don't have the vocabulary to understand the
10152        terms and what they are trying to do. I'll get there
10153 
10154        what is the best definition for 'mold' ?
10155 
10156        nm found them in 2.3.2 will figure it out
10157 
10158~rabsef I’m not really sure I fully get mold yet tbh
10159 
10160~somlus got it - going through the list of atoms exercise, ^- functions
10161        to typecast by mold.
10162 
10163        that's where I got to looking for it.
10164 
10165        Question - in the list walkthrough, the null value (line 6) is
10166        the next line after the logic test "does end equal count" - does
10167        this null end the generator because it is a null, or does it add
10168        the final null onto the list, and this then ends the generator
10169        because the ^- rune created a typcast by mold command, which
10170        defined the mold as a list of atoms and as a list, needed to be
10171        null terminated, and with this final null the list is complete
10172        so that command is complete, so the generator is complete? I
10173        hope that makes sense. wondering if ~ is basically one of the
10174        commands to terminate a generator, or simply initiated a cascade
10175        of completion which finished all the work the generator knew it
10176        had to do
10177 
10178~rabsef it makes sense
10179 
10180        the null ends the list because it is a final value for the duple
10181        cell being created each recursion (I’m not looking at it but
10182        that’s what I remember)
10183 
10184        null does not automatically end a cell/list but for needing to
10185        be the terminal tail of a list but adding one to the first or
10186        tail-most-tail position of a list doesn’t automatically end what
10187        ever you’re building. you could add another element to the right
10188        and screw it up as a list but still have a valid cell I believe
10189 
10190        your latter wondering is correct
10191 
10192~somlus thanks
10193 
10194        In the rune explanations, there is a regular syntax and an "AST"
10195        Abstract Syntax Tree syntax. Why?
10196 
10197        What's the value of having two ways of describing syntax?
10198 
10199~rabsef that’s a good one and idk. here is an experiment for ur last
10200        question. change cast to list to cast to cell (kethep ket) and
10201        change the null to ‘end’ and see the cell it produces
10202 
10203~somlus end in single quotes?
10204 
10205        getting only errors on that experiment
10206 
10207        this is what I did
10208 
10209        https://pastebin.com/snvBDACx
10210 
10211        we can forget about it - I'm satisfied that i know what's going
10212        on with the end of that generator
10213 
10214        New question:
10215 
10216        https://pastebin.com/jarRwn4X
10217 
10218        in this example, the list ends in the number 100. The question
10219        is: how does the ?: know where to send the "false" output
10220        computer next. I added :- 100 but it did not consider that to be
10221        the first and ~ to be the second child. it knew that :- 100 and
10222        ~ were the first child and :- count onward was the 2nd child.
10223        How did it know that without any confusion?
10224 
10225~rabsef sorry for my belated reply - it's the wrong kind of single
10226        quotes in there
10227 
10228        ' ' not ` `
10229 
10230        I typed it wrong sorry
10231 
10232        It's the other way. Each :- for the no condition opens a cell
10233        and gets a left hand value [1 [2 [3 [4 ~]]]] or [1 [2 [3 [4 [100
10234        ~]]]]]
10235 
10236        in the next recursion, then, the no condition again opens a cell
10237        :-(1 :-(2 :-(3 :-(4 {yes condition} ~))))
10238 
10239        look at all those unhappy faces
10240 
10241        :-(1 :-(2 :-(3 :-(4 ~))))
10242 
10243        `(list @ud)`:-(1 :-(2 :-(3 :-(4 ~))))
10244 
10245        It might help to write that one down - each recursion write out
10246        the hoon that is happening and what is left open
10247 
10248        :- has 2 children, is given one before buck so it's still open -
10249        so buck calls back to the |- trap /inside/ of the second child
10250        of the first :- and so on until ~
10251 
10252        `(list cord)`:-('countprime' :-('recursionone' :-('recursiontwo'
10253        :-('recursionthree' ~))))
10254 
10255        :-('countprime' :-('recursionone' :-('recursiontwo'
10256        :-('recursionthree' ~))))
10257 
10258        ['countprime' ['recursionone' ['recursiontwo' ['recursionthree'
10259        ~]]]]
10260 
10261        Does that make sense?
10262 
10263~somlus studying it... (thanks BTW)
10264 
10265~rabsef my pleasure - let me know if you want to go more on it
10266 
10267~somlus Yes, makes sense. That is how I understood it. So the final null
10268        slams shut all of the cells. it closes the right most cell, and
10269        that makes the right most part of the cell to the left of that
10270        one and so one, ok great. thanks
10271 
10272~rabsef just to be clear it only 'closes them' because it is a value
10273        (arm) rather than more computation (leg)
10274 
10275        and :- needs 2 arms [one two]
10276 
10277        I think you get that but it's not about that value being ~
10278 
10279~somlus oh I thought it closed them because the generator specified that
10280        the final product needs to be a list, ending in null, so once
10281        you end in null, the generator says "oh I'm done - I built what
10282        I said I would build"
10283 
10284~rabsef no - suppose you cast as a list but instead of ~ put 'end' -
10285        you'll get a nest fail
10286 
10287        because... it's not a list because the last term is 'end'
10288 
10289        but it is a cell
10290 
10291        :-(1 :-(2 :-(3 :-(4 'end'))))
10292 
10293        `(list @ud)`:-(1 :-(2 :-(3 :-(4 'end'))))
10294 
10295        `(list @ud)`:-(1 :-(2 :-(3 :-(4 ~))))
10296 
10297        Perhaps u could use some more LSD, doc?
10298 
10299~somlus haha! yeah, send it over.
10300 
10301        trying now to rewrite it as you suggested earlier, to not
10302        require it to be a list of numbers, but instead a list of nouns
10303 
10304~rabsef wait tho - list means one thing in specific
10305 
10306        ['a' ['series' ['of' ['nested' ['cells' ['that' ['are'
10307        ['null-terminated' ~]]]]]]]]
10308 
10309~somlus sorry you said cast to cell.
10310 
10311        ok trying that
10312 
10313~rabsef `(list cord)`['a' ['series' ['of' ['nested' ['cells' ['that'
10314        ['are' ['null-terminated' ~]]]]]]]
10315 
10316        ['a' ['series' ['of' ['nested' ['cells' ['that' ['are'
10317        ['null-terminated' ~]]]]]]]
10318 
10319        ['a' ['series' ['of' ['nested' ['cells' ['that' ['are'
10320        ['null-terminated' ~]]]]]]]]
10321 
10322~somlus yes sir. can't end in 'end' or it's not a list
10323 
10324~rabsef sorry that took me a min lol
10325 
10326        `(list cord)`['a' ['series' ['of' ['nested' ['cells' ['that'
10327        ['are' ['null-terminated' ~]]]]]]]]
10328 
10329~somlus u, sir are committed
10330 
10331~rabsef you should see all the garbage executions that pile up at the
10332        bottom of my screen lol
10333 
10334        this is why I cannot have nice things
10335 
10336        although 10.3 is hella stable on this new OTA
10337 
10338        I haven't crashed in over a day
10339 
10340~risruc important point ^ rab. it doesn't close all the cells because it
10341        returns ~, it closes the cells because it returns an atom, not
10342        :-
10343 
10344~rabsef ^!
10345 
10346~somlus thanks
10347 
10348~rabsef that is the /ultimate/ point
10349 
10350        not to back pat myself
10351 
10352~somlus what am I missing here?
10353 
10354        https://pastebin.com/UDAYBGWc
10355 
10356        trying to convert this to a non-list cell output. used ket for
10357        the mold
10358 
10359~rabsef does it error?
10360 
10361~somlus and put the cord 'end' in there
10362 
10363        yes
10364 
10365~rabsef just comment out the kethep
10366 
10367        mabe you can't cast as ^
10368 
10369~somlus that worked
10370 
10371~rabsef my b
10372 
10373~somlus ok cool.
10374 
10375~rabsef u could ^- [* *]
10376 
10377        do you understand why that would work? - think about it
10378 
10379~somlus [1 [2 [3 [4 'end']]]]
10380 
10381~rabsef relates to subject
10382 
10383~somlus that's the output neat
10384 
10385        gimme a minute to consider
10386 
10387~rabsef yah - sorry that's what I wanted you to get earlier lol - I was
10388        writing from fone
10389 
10390        i'm kinda surprised u can't cast as ^ cuz you can cast as @
10391 
10392        seems weird
10393 
10394~somlus if everything is a cell, you don't need to state that it is a
10395        cell. that is the status quo
10396 
10397~rabsef sort of - If * means anything, an @ or a ^ then [* *] can be [1
10398        2] or [1 [2 [3 4]]]
10399 
10400        in the second example, the first * is 1 and the second is [2 [3
10401        4]]
10402 
10403~risruc you could as well just use * :)
10404 
10405~rabsef that is true lol
10406 
10407        which is why you should be able to ^
10408 
10409~somlus dinner's ready heading afk a bit
10410 
10411~rabsef I might not get why ^ doesn't work there
10412 
10413~somlus thx for the help though
10414 
10415~rabsef my pleasure - let's do some more later if I'm around :)
10416 
10417~somlus ok cool if you're game. I figured you were learning on yourown
10418 
10419~rabsef I am but I'm stuck so I'm happy to think w/ you
10420 
10421~somlus ok, here's another Q:
10422 
10423        Pastebin to follow question
10424 
10425        in this example, the list ends in the number 100. The question
10426        is: how does the ?: know where to send the "false" output
10427        computer next. I added :- 100 but it did not consider that to be
10428        the first and ~ to be the second child. it knew that :- 100 and
10429        ~ were the first child and :- count onward was the 2nd child.
10430        How did it know that without any confusion?
10431 
10432        https://pastebin.com/jarRwn4X
10433 
10434~rabsef https://hastebin.com/kuhayucuvu.shell
10435 
10436        I'm not sure what you mean exactly. The /last/ value of your
10437        cell opened in a yes condition is just a cell
10438 
10439        the cell of 100 ~
10440 
10441~risruc the ~ is the last entry in the list. in case it's confusing that
10442        it displays it first in dojo.
10443 
10444~rabsef the first child of the first cell is the furthest left value
10445 
10446        ahhh cuz list
10447 
10448        my b - didn't understand wat u meant
10449 
10450~risruc if you run that with +gen 10
10451 
10452        you'll get ~[1 2 3 4 5 6 7 8 9 100]
10453 
10454        which of course is [1 2 3 4 5 6 7 8 9 100 ~]
10455 
10456~rabsef or [1 [2 [.... [100 ~]...]]
10457 
10458~somlus So the boolean logic asks if end equals count. if not so then it
10459        goes to the next line that say "add 100 to the list" and then I
10460        would think it would say "ok so that's what we do if true, so
10461        the next line is what we do if false, and then it would end on ~
10462        but it does not. It knows that both 100 and ~ are what you do if
10463        true. How does it know clearly where the "if false" part begins?
10464 
10465        hope this makes sense.
10466 
10467~risruc [1 [2 [3 [4 [5 [6 [7 [8 [9 [100 ~]]]]]]]]]
10468 
10469~rabsef Ah no - because you start the "true" line w/ :-
10470 
10471~somlus I figured the true/false routing was by line, but it is clearly
10472        not. so if not, then what is it by?
10473 
10474        oh shit that easy?
10475 
10476~rabsef so it says "oh I gotta make a cell first value of... 100 ...
10477        second of ... ~"
10478 
10479        because :- has how many children? Let me hear it from the back
10480        this time
10481 
10482~somlus that's why hoontards....
10483 
10484        three - the logic test, the thing if true and the thing if false
10485 
10486~rabsef effectively you're just skipping 11-99 hehe
10487 
10488~somlus yeah I hate 11-99
10489 
10490~rabsef that's ?: - :- has 2 children value 1 and value 2
10491 
10492        or noun 1 and noun 2
10493 
10494        :-([1 [2 [4 5]]] 6)
10495 
10496~somlus oh ok. but the syntax is still (logic test) (if true) (if false)
10497 
10498~rabsef :- 2 ~
10499 
10500~somlus I may not have noticed that ?: always has :- to trigger the if
10501        false case....
10502 
10503~risruc :- 2 ~
10504 
10505~rabsef wait let's slow down
10506 
10507~risruc :- 2 ~
10508 
10509~rabsef that's not quite right
10510 
10511~risruc :- 2 ~
10512 
10513~rabsef The TRUE condition of ?: is "some hoon" as they are fond of
10514        saying
10515 
10516~risruc lol. thid time's a charm
10517 
10518~rabsef hoon can either be a value (~ e.g.), or another program
10519 
10520        in the exemplary presentation, it is the value ~
10521 
10522        in your modified presentation, it is the program :-
10523 
10524~somlus yeah I'm looking at the wutcol hoon expression now and :- is not
10525        mentioned.
10526 
10527~rabsef that program requires 2 children to close
10528 
10529        so - before you put in :- 100 ~, it closed instantly because ~
10530        is a leg/value/not a program
10531 
10532        when you put in :-, EVERYTHING after :- is part of the true
10533        condition until :- is satisfied (2 children or 100 and ~)
10534 
10535~somlus oh, yeah. I forgot the naivety of needing to be certain when
10536        things are done.
10537 
10538~rabsef the most important dunce-ing
10539 
10540        of the hoonening
10541 
10542~somlus yeah
10543 
10544        So when you look at code are you trying to zone in on where
10545        values are determined, and then you know "ok these are the
10546        endpoints of this generator"?
10547 
10548~rabsef Let's say it even a little differently - each time you see a
10549        rune you have to determine where it 'goes'
10550 
10551~risruc you must become one with the interpretter.
10552 
10553~rabsef that is to say when it ends
10554 
10555        as ~risruc says
10556 
10557~risruc haha
10558 
10559~somlus i am not one with the interpreter, to be clear.
10560 
10561~rabsef you will be
10562 
10563        it becomes more natural.. Here is a story I will tell:
10564 
10565        When I was a kid, I couldn't type. My mom got me Mavis Beacon
10566        teaches typing, Mario teaches typing etc
10567 
10568        I hated them
10569 
10570        Then I found this MMORPG MUD game called Castle Marrach online
10571        one summer
10572 
10573        you had to type like "go north" "don cap" etc - it was a text
10574        game
10575 
10576        just like zork only many people
10577 
10578        so you had to type fast to make your character do things in a
10579        convincing way when read out on the screen right?
10580 
10581~somlus sounds cool
10582 
10583~rabsef you don't want to be slow to "accept cheese from alice" in the
10584        cafeteria
10585 
10586~somlus and there it was -> motivation
10587 
10588~rabsef it was the best game - I go back every once in a while lol
10589 
10590        yes well so I learned how to type
10591 
10592        but further, I truly thought in the parser language for a brief
10593        moment
10594 
10595        like I would be outside and think "take box from ground" etc
10596 
10597~somlus haha yes that makes sense.
10598 
10599~rabsef it just became so natural to think in that parsing cuz I played
10600        it non stop for like 3 months
10601 
10602~somlus Well I can say that I am definitely enjoying myself, this whole
10603        thing is like a gigantic puzzle, an enormous brain teaser (is
10604        that the phrase)?
10605 
10606~rabsef an enigma wrapped in a conundrum
10607 
10608~risruc haha
10609 
10610~rabsef w/ a side of mystery
10611 
10612~somlus I chisel off a chunk of minor understanding from the whole and
10613        then that lets me go back to something that was greek and then
10614        it's a tiny bit less greek, where I can chisel off another small
10615        piece of understanding
10616 
10617~rabsef that's literally how I feel every step of the way - I just have
10618        more runway than you do rn
10619 
10620        /have traversed more runway
10621 
10622~banhex i loved mavis beacon and still think about it from time to time
10623 
10624~rabsef you should try castle marrach
10625 
10626        I wish that game were still popular
10627 
10628        there are like only 5-10 people online at any given time now. In
10629        it's heyday it was like 40-50
10630 
10631~risruc lol.....
10632 
10633~rabsef and I played an artist and they custom coded me a paint set
10634        object I could interact w/
10635 
10636        it was so sick.
10637 
10638~somlus it sounds awesome. I literally played zork first on a green
10639        screen monitor on the 2nd ever IBM PC
10640 
10641~rabsef blessed
10642 
10643~somlus I was already 872 years old at that time.
10644 
10645~rabsef I used to make christmas trees out of x's on the terminal
10646        screens at my moms office when I was sick from school
10647 
10648        I get sick a lot (or did years ago)
10649 
10650~somlus yeah, I remember the computers at dad's office (IBM) - so
10651        alluring and confusing.
10652 
10653        I did too. I was so sick they diagnosed me with cystic fibrosis.
10654 
10655        In error, fortunately
10656 
10657~risruc :- 3 :- 2 :- 1 ~
10658 
10659~rabsef ~ris - do my examples print for u? That one didn't for me
10660 
10661~risruc that help ~som?
10662 
10663        oh.
10664 
10665        haha
10666 
10667~rabsef :- 3 :- 2 :- 1 ~
10668 
10669~risruc sometimes yes, sometimes no.
10670 
10671~rabsef weird right?
10672 
10673~risruc well, it produced [3 2 1 ~]
10674 
10675~rabsef I see it on my end lol
10676 
10677~somlus I did not see the result, just the code that leads to the
10678        result.
10679 
10680        But it's good to see that :- can be done in succession to arrive
10681        at nested cells.
10682 
10683~risruc thought it might help ~som understand
10684 
10685        to see what is actually being done written out on one line
10686 
10687        that's precisely what your generator does.
10688 
10689~somlus yeah it does actually.
10690 
10691~rabsef if one of our examples doesn't print just repeat it w/ # on ur
10692        end
10693 
10694        cuz we're probably trying to show off and it's frustrating if u
10695        cant see it
10696 
10697        hehehe
10698 
10699~somlus oh cool i'm reading you can :- more than once
10700 
10701~risruc the reason it ends is because the last entry is not :-
10702 
10703~rabsef Yes - why?
10704 
10705~risruc not because the last entry is ~
10706 
10707~rabsef or ris u stepped on it
10708 
10709        oh*
10710 
10711        :)
10712 
10713~risruc haha
10714 
10715        ya i blew it.
10716 
10717~rabsef say it in ur own words ~somlus - why can you chain :-
10718 
10719~somlus because it arrives at an endpoint which is a noun, and a cell
10720        (actually nested cells) and not something that say "hey i've got
10721        something else for you to do)
10722 
10723        that's why it ends
10724 
10725~rabsef >(actually nested cell) - blessed understanding shines on u this
10726        day
10727 
10728        now consider re: arm and leg
10729 
10730        :-(leg leg) terminates instantly
10731 
10732~somlus and why it can chain is because it builds a cell using the right
10733        two nouns (in this case atoms) after the right most :- and then
10734        that becomes the right most noun for the atom and :- to the left
10735        of that]
10736 
10737~rabsef :-(leg arm) has to resolve arm
10738 
10739~risruc yep, you got it.
10740 
10741~somlus ah yeah, basically super important to keep thinking in terms of
10742        "this has to resolve before..."
10743 
10744~risruc once the last child has 2 args, it closes and becomes the second
10745        arg for the penultimate :-, which closes and becomes the arg for
10746        the antepenutimate :-, et c.
10747 
10748~rabsef now consider the subject - a giant binary tree where some
10749        positions are arms and some are legs - our naked gen (any naked
10750        gen) is an arm branch that requires a leg being sent to it
10751        (sample) and then resolves itself
10752 
10753        >antipenultimate :D
10754 
10755        ante*
10756 
10757~somlus yeah, I actually understand that now. I'm sort of blown away /
10758        in awe that we're in the middle of a giant tree right now and
10759        the more fundamental nodes, living by the same rules, are above
10760        us on the way to the core node, nock
10761 
10762        again, please pass the LSD
10763 
10764~rabsef let us consider and meditate for a moment on the mind that
10765        envisioned this
10766 
10767        Curtis is a fugging galaxy brain
10768 
10769~somlus totally.
10770 
10771~rabsef It's downright implausible that a mind could think of this, in
10772        my opinion
10773 
10774~risruc btw somlus, that generator ending with 100 ~ is adorable
10775 
10776~somlus thanks man, I named him ~george
10777 
10778~rabsef +george 99, the perfect proram
10779 
10780~risruc [1 2 3 100 ~]
10781 
10782        lol!
10783 
10784~somlus Yeah, it's how I play hide and go seek count to 100.
10785 
10786        +george 5
10787 
10788~rabsef lol
10789 
10790        Adam Carolla has this bit "You know it's funny when people say
10791        'You can run but you can't hide'. Of course they can hide.
10792        Really the saying should be 'You can run, and you can hide, but
10793        I will find you anyway'"
10794 
10795~somlus yes perhaps the word "forever" got lost off of that saying in
10796        the middle ages
10797 
10798~risruc man that just cracked me up some.
10799 
10800        som.
10801 
10802        wife "hm?"
10803 
10804        > nerd things
10805 
10806        no further questions
10807 
10808        to get great
10809 
10810~somlus totally. I just showed my wife the +list generator and it's full
10811        output and she was like. "ummmmmmm..... you've been in that
10812        chair since 5am you know?"
10813 
10814        I'm like "yeah but I'm happy"
10815 
10816~risruc to her great credit, she's learned to take that as a good enough
10817        answer.
10818 
10819        haha!
10820 
10821~somlus indeed
10822 
10823~risruc yeah man. learning is the bomb.
10824 
10825        and this feels special
10826 
10827~rabsef LOL
10828 
10829        My fiancee is always like "So how's your... urbit stuff?"
10830 
10831        I love it that they're supportive but also not invested - it is
10832        cute as hell
10833 
10834~somlus yeah it's nice. I think that chicks dig it when their men are
10835        super into super harmless stuff. for lots of the right reasons,
10836        plus they can turn off their brains to worrying about the man's
10837        well-being or whether he is trending toward up to no good.
10838 
10839~rabsef lol fiancee found out about CY founder of tlon today and wiki'd
10840        him and we did have to have a multi-part discussion about
10841        why the wiki on him is grossly inaccurate and abusive to his
10842        character
10843 
10844~somlus yeah, much to discuss there...
10845 
10846        might help the project actually - all news is good news in a way
10847        to a project that wants to be known
10848 
10849~rabsef It's frustrating because I read UR and think it's clearly the
10850        work of a genius and I have a major issue w/ constraining genius
10851 
10852        Like - just let these people think and produce because whatever
10853        it is it will be forward
10854 
10855        more than now, which is objectively good
10856 
10857~somlus Good morning guys. I realize we are not due to being for another
10858        40 mins. Wanted to set some expectations. I am still all-in as
10859        always. Having said that, all 12 hours of yesterday's work were
10860        basically re-learning, organizing notes, solidifying and not
10861        moving beyond a generator unless I pretty much fully undersood
10862        it. That took some time, since some concepts are abstract and
10863        needed to work through those.
10864 
10865        As such, I have not studied the Caesar Cypher, but even if
10866        unprepared I know that I will learn a lot so I am still ready to
10867        jump into the pool of molten lava for another Hoon baptism.
10868 
10869        *not due to meet
10870 
10871~nibtyc Here as moon
10872 
10873~tomsef hi fellow hoontards
10874 
10875~rabsef hey good morning.
10876 
10877        we will be starting shortly. somlus, I want to spend the first
10878        half just looking at the structure and learning how to read it
10879        so we can then dissect the arms thereafter, and in order
10880 
10881        so no need to be super prepared
10882 
10883~somlus ok
10884 
10885        those were the magic words today.
10886 
10887~rabsef I’ll be on in 4. just taking care of. few things and getting out
10888        my Sunday best.
10889 
10890~somlus take your time, i'm chilling and getting everything in its place
10891 
10892        cyper loaded and working in moon
10893 
10894        cypher
10895 
10896        hey tomsef!
10897 
10898        might not chat as moon so that I can have the moon screens all
10899        be full of vscode and dojo
10900 
10901        we'll see
10902 
10903        tomsef, we are going to go over the caesar cypher from section
10904        1.7.1 of the walkthrough. You probably saw that but wanted to
10905        make it clear
10906 
10907~rabsef I arrive!
10908 
10909        let me pull that bad boy up r quick
10910 
10911        Ok - so everyone here who wants to be here?
10912 
10913~somlus I reckon
10914 
10915~rabsef Last time we went over a little bit of the introductory material
10916        for caesar. I want to return to that and go through the first 6
10917        lines carefully again
10918 
10919~somlus great
10920 
10921~rabsef They will help us in several ways, (1) we should look at how
10922        gates are called and what they require (2) we should look at the
10923        function of =. and what it's doing and (3) we should use this
10924        opportunity to reinforce our discussion of :- from yesterday
10925 
10926~somlus ~tomsef don't be shy with questions
10927 
10928~rabsef https://urbit.org/docs/tutorials/hoon/caesar/
10929 
10930        ^this is what I'm talking about incidentally, if you want to
10931        follow along
10932 
10933        Line 2 opens a gate - what does it expect?
10934 
10935        it calls a gate I guess - it doesn't open it; bad phrasing
10936 
10937~somlus tape and...
10938 
10939~rabsef yes two values, in a cell, a tape and a @ud or an unsigned
10940        decimal
10941 
10942~somlus an input from the user which is the tape and a real number
10943 
10944        which is the key
10945 
10946        *which is a tape
10947 
10948~rabsef We should note that, in every arm beneath the |% creation of a
10949        core, the arms begin with gates
10950 
10951        the gates require various information points - some 2 items,
10952        some more. This is going to be important for us.
10953 
10954~somlus ok
10955 
10956~rabsef save for alpha which is basically a leg of data
10957 
10958        So - when we get further into this and we see an arm get called,
10959        we should be able to intuit from looking at that arm what data
10960        structures the arguments are (arm arg arg2 arg3) of arm |= @ud
10961        @tas tape would require (arm 1 %term "here")
10962 
10963~somlus ok
10964 
10965        I get it so far
10966 
10967~rabsef sometimes its easiest to think of it that way so you can
10968        remember what the arguments you're calling are and better
10969        understand how they're getting there, especially when an
10970        argument is a call to an arm itself (see (operate message key
10971        (encoder key)) where (encoder key) is a call to the encoder arm
10972        and requires an @ud
10973 
10974~somlus ok
10975 
10976~rabsef aight - line 3 "composes two expressions, inverted"
10977 
10978        what the hell does that mean?
10979 
10980        also, for bonus points, identify the two expressions here
10981 
10982~somlus one moment
10983 
10984        thinking...
10985 
10986~rabsef Here is a hint: It is proper hoon style to keep the "heaviest"
10987        part of the program towards the bottom - the top should be
10988        reserved to show you what's going on as output. This style
10989        choice correlates to what we're asking
10990 
10991        heavy = more hoon basically
10992 
10993~somlus the shift and unshifted messages
10994 
10995~rabsef those are what? the two expressions? one each?
10996 
10997~somlus and just before that you say "convert this message to lower
10998        case"
10999 
11000        one moment
11001 
11002~rabsef (the proposed two expressions one each I just made is incorrect
11003        - think of the hint)
11004 
11005~somlus ok
11006 
11007~rabsef It's pretty hard - let's work through it together? Or you want
11008        another sec?
11009 
11010~somlus hmm sorry I keep concluding that you are outputing an upshift
11011        in characters and also a downshift in characters based on the
11012        number inputted. I do not see output beyond that
11013 
11014        yeah let's go through it.
11015 
11016        there is a shift and an unshift
11017 
11018~rabsef =<(output [possibleoutput=1 output=2 otherpossibleoutput=3])
11019 
11020~somlus and then the stuff that supports that
11021 
11022~rabsef You're /close/
11023 
11024        Yes exactly that's right but perhaps not how I'd phrase it
11025 
11026        The expressions in lines 4-6 are the first expression of hoon,
11027        and everything thereafter is the second expression of hoon
11028 
11029        the first expression is evaluated in light of the second
11030 
11031        =<(output [possibleoutput=1 output=2 otherpossibleoutput=3])
11032 
11033~somlus oh, yes ok got it.
11034 
11035~rabsef If I just "output" it would be like "find.output"
11036 
11037        because output is not in ... the subject
11038 
11039~somlus "I need this" (and then hoon thinks, "ok so in order to get that
11040        then let's go tackle this that will get us the thing requested")
11041 
11042        and each is an expression
11043 
11044~ribnys hoon schooler hanging out here, at least intermittently
11045 
11046~somlus ok that concept makes sense now. expression is sort of = bundle
11047        of work that needs to reach completion in order for the next
11048        bundle of work (or solution) can be achieved.
11049 
11050~rabsef that's right =< points lines 4-6 to the core line 7 until close
11051        as "the other expression to evaluate the first, in light of"
11052 
11053~somlus cool ribnys, welcome. please don't let our fast pace hold you
11054        back from asking any questions. rab and I are pretty comfortable
11055        with one another.
11056 
11057~rabsef yes exactly - a hoon expression is something that ends at some
11058        point. The core ends at some point (the --) and the lines 4-6
11059        end at some point
11060 
11061        now this brings us to point 3 from earlier - how do we know
11062        lines 4-6 end at some point?
11063 
11064        Yah I'm happy to remediate on any part here.
11065 
11066~somlus because a command has stated that they determine an input.
11067        I'm looking through runes now. it's either =< or :- I think.
11068        checking
11069 
11070        =<
11071 
11072~rabsef Just as with our discussion of recursion, where :- has 2
11073        children and terminates producing a cell, on line 6, we have the
11074        last child of :- so the output of 4-6 is a cell of (shift msg
11075        key) (unshift msg key)
11076 
11077~somlus compose two nouns, inverted - that is what the generator need to
11078        do before it can go back to bed...
11079 
11080~rabsef once those arms (shift... and unshift...) are resolved to
11081        values, the :- can output on screen
11082 
11083~somlus oh maybe not.
11084 
11085~rabsef that's right I guess - it is composing two nouns and evaluating
11086        the first in light of the second
11087 
11088        no you're right it's just not the point I was trying to drive at
11089 
11090~somlus let's actually make sure.
11091 
11092~rabsef the way we can determine where one noun ends is when it
11093        naturally would make out a product
11094 
11095~somlus :- is the most specific command - "construct a cell"
11096 
11097~rabsef lines 4-6 make out a product when they (1) mutate the subject
11098        and then (2) construct a cell (a duple)
11099 
11100~somlus Since I do not fully understand the importance of =< I jumped
11101        the gun there.
11102 
11103        ok
11104 
11105~rabsef You're circling the importance. In order to understand =< we
11106        have to be able to interpret what the two expressions are
11107 
11108        with something like:
11109 
11110        =<(output [possibleoutput=1 output=2 otherpossibleoutput=3])
11111 
11112        it's easy - it's output and the cell (1 and 2 respectively)
11113 
11114        but with a larger program it is more difficult - so you need to
11115        figure out where one expression is 'complete'
11116 
11117        lines 4-6 are a complete thought - they do not 'keep going'
11118 
11119        like the ~ at the end of the recursion
11120 
11121        we know that we've completed the work
11122 
11123        does that make sense?
11124 
11125~somlus ok. can you tell me where this notation above came from that
11126        begins with =<(out.... and ends with 3]) ?
11127 
11128        I cannot tell where you got that / where it came from
11129 
11130~rabsef that's just an example of =< being used in a very simple context
11131 
11132        with =<(output [possibleoutput=1 output=2
11133        otherpossibleoutput=3]) we are saying "evaluate output in light
11134        of our second expression of hoon, which is a cell that gives
11135        output a value"
11136 
11137        and it returns that value, 2, in light of that cell, when output
11138        is evaluated
11139 
11140~somlus To your point where this is like the ~ in the list, yes I
11141        totally understand that. You tell Hoon what you are going to do
11142        and then once you have clearly done that, Hoon knows to stop.
11143 
11144~rabsef since output is just a face, we know we are done with output
11145        when we produce the value that face denotes
11146 
11147~somlus one moment pls...
11148 
11149~rabsef Let me give one more example:
11150 
11151        =<(producttwo [productone=(add 1 1) producttwo=(add 2 2)])
11152 
11153        but see:
11154 
11155        producttwo
11156 
11157~somlus ok
11158 
11159        that I do understand
11160 
11161~rabsef without =< in front of producttwo, the computer does not know
11162        what I'm referring to
11163 
11164        with =< it knows to look at /the second expression of hoon/ to
11165        find how to evaluate the first
11166 
11167~somlus ok
11168 
11169~rabsef Here is a tricky part -> The values established at the gate are
11170        /also/ part of what can be evaluated in the first expression
11171        after =<
11172 
11173~somlus clear now
11174 
11175~rabsef That is to say "msg" and "key" are still available to the first
11176        expression for evaluation
11177 
11178        so /basically/ =< adds the second expression to the context
11179        available to the first expression for evaluation
11180 
11181~somlus when you say first expression here you mean the first one to
11182        occur, which is lower in the code correct?
11183 
11184~rabsef sorry no. The first expression is linearly the one that comes
11185        first
11186 
11187        =<(producttwo [productone=(add 1 1) producttwo=(add 2 2)])
11188 
11189~somlus ok got it, realized that from your last line
11190 
11191        yes
11192 
11193~rabsef "producttwo" by itself is the first expression; the cell is the
11194        second expression
11195 
11196~somlus copy that
11197 
11198~rabsef here lines 4-6 are "the first expression" and lines 7-n are "the
11199        second expression"
11200 
11201~somlus ok, I'm with you so far.
11202 
11203~rabsef and what I want to establish most is that we have to be able to
11204        determine where an expression terminates to be able to see first
11205        from second
11206 
11207        it's sort of a pedantic point but it's critical for reading hoon
11208        imo
11209 
11210~somlus agreed.
11211 
11212        and yes I now do see that.
11213 
11214~rabsef ok so @ line 2, we have a sample of a tape and an @ud and that's
11215        all we really have to work with
11216 
11217        So w/ JUST that information we could not complete lines 4-6
11218 
11219        but line 3 comes along w/ =< and it adds to the context
11220        available for evaluating 4-6, 7-n below it.
11221 
11222        now /both/ line 2 and 7-n are available to help evaluate 4-6
11223 
11224~somlus so is =< basically a message saying "don't freak out, there is
11225        more down below, put a placeholder here and what follows is
11226        going to make this make sense"?
11227 
11228~rabsef Last reiteration of this -> if we comment out line 3, we would
11229        error on line 5 with find.shift
11230 
11231~somlus I know that's juvenile, but I always try plain english where I
11232        can
11233 
11234~rabsef which just means the compiler doesn't know wtf shift is
11235 
11236        yes - that's it
11237 
11238~somlus great
11239 
11240~rabsef ok - so, now we can move forward
11241 
11242        lol - sorry
11243 
11244~somlus cool
11245 
11246~rabsef Ok so line 4 is important too, but you've already said what
11247        it does =. creates a 'mutated' sample where a change has been
11248        made to it and the rest of the program is evaluated off of that
11249        mutation
11250 
11251~ronlyr just want to say, this is a great walkthrough... I'm just
11252        lurking and following along.
11253 
11254~rabsef There's some significant nuance around what is actually
11255        happening w/ the subject there but I don't really think we need
11256        to get into it, and I'm not sure I understand anyway
11257 
11258        Thanks ~ronlyr :)
11259 
11260~somlus cool ronlyr! I was actually wondering whether it had lurker
11261        value
11262 
11263~rabsef so for our purposes msg is converted using (cass) into a new
11264        form of msg
11265 
11266~minder It does
11267 
11268~rabsef cass makes a tape lower case
11269 
11270~somlus ok. also rab, I want to get back to the "this is the tricky
11271        part" you mentioned above when we're done
11272 
11273~rabsef Oh hit me
11274 
11275~somlus I did not see anything as tricky, but I will
11276 
11277        no let's wait until the end. it will have more context
11278 
11279~rabsef yes ok
11280 
11281~somlus that's to remind us
11282 
11283        keep on keepin' on
11284 
11285        ~minder's here! that man is busy if he says it has value, then
11286        it does
11287 
11288~rabsef Ok so cass makes a tape lower, cuss makes it upper - so =. msg
11289        (cass msg) takes our incoming "TAPE" or "TaPe" and makes it,
11290        regardless of input format "tape"
11291 
11292~somlus I learned that from your ingeniuous spongebob sarcastic text
11293        converter.
11294 
11295~rabsef incidentally cass and cuss are part of the standard library and
11296        easy to find using / in the documentation - for those who wonder
11297        how you are to know that
11298 
11299        lol thanks :)
11300 
11301        why might it be important to have everything lower case?
11302 
11303        Well we know the function of this program - it's going to take
11304        in a message and shift it by a given amt
11305 
11306        and we can look at line 8 and see that the tape we're going to
11307        use to shift - all lower case values
11308 
11309        and:
11310 
11311~somlus yeah, and you don't want to reveal any of the encoded text by
11312        telling anyone which characters started out as caps
11313 
11314~rabsef ?:(=("a" "A") %.y %.n)
11315 
11316        that's not quite why - it's because the generator is too naive
11317        to handle different casing
11318 
11319        later, we will make a map where something like a=b, b=c, c=d is
11320        going on in a big list of all letters - if our msg value had
11321        upper case, it would fail for finding the letter to look for
11322 
11323~somlus true, you could make it that way. but also as a cypher its' dumb
11324        to preserve casing and tell them germans that you're going to
11325        bomb Xuxenberg
11326 
11327        *m
11328 
11329~rabsef "A" is not "a" for the compiler, and so if we only have a map
11330        of lower case letters, we have to lower case our input for the
11331        program to work
11332 
11333~somlus yes all true
11334 
11335~rabsef lol
11336 
11337        that is also true
11338 
11339        Xuxenberg is a lovely country
11340 
11341~somlus As an extra task we might consider converting this to one that
11342        can handle case, some other day.
11343 
11344~rabsef it is imminently doable
11345 
11346        would make everything a little longer tho
11347 
11348~somlus ok continuing...
11349 
11350~rabsef ok - so lines 5 and 6 create a cell
11351 
11352~somlus yes
11353 
11354~rabsef the cell is of two calls to arms below using the values we have
11355        in our context
11356 
11357        so here's where we start looking at the arms, and it's going to
11358        get hectic
11359 
11360        let's focus just on shift
11361 
11362~somlus col
11363 
11364~rabsef Shift requires a tape and an @ud, by it's gate - so we pass a
11365        tape and an @ud - namely msg and key
11366 
11367~somlus cool
11368 
11369~rabsef that's the notation there (arm argument argument argument)
11370 
11371        pass means to give the argument to the arm so that it may
11372        compute, to restate that
11373 
11374~somlus question why above did we five it the face msg and now we are
11375        calling it message
11376 
11377~rabsef that's sort of a tricky question - you could call it msg again,
11378        but then you have two values w/ a face of msg in the subject
11379 
11380        the compiler finds the first one and uses that when it's
11381        thinking - you can use other notation to tell it to skip the
11382        first one it finds
11383 
11384~somlus oh it's a different value, I should have been more patient.
11385 
11386~rabsef it wouldn't really matter here, but consider that within arms it
11387        might be good to keep different names so you're being explicit
11388        w/ the value you want to call
11389 
11390~somlus I assumed it was the same thing, the user input.
11391 
11392~rabsef it's not different exactly. At line 10, when the msg is passed
11393        to Shift as message, message = msg = message
11394 
11395~digmev Hey - so i have some context, what are you guys talking about?
11396 
11397~somlus wait, is this like when your spongebob generator built a tape as
11398        it tore down another tape character by character?
11399 
11400        hi divmev
11401 
11402~rabsef the only difference is that the same value "our message" has
11403        a face of msg and a face of message. Suppose you transformed
11404        message in shift and then later wanted to call it - if you had
11405        named it msg, it would be a pain in the ass because you have msg
11406        qua original and msg qua modified
11407 
11408~somlus digmev
11409 
11410        https://urbit.org/docs/tutorials/hoon/caesar/
11411 
11412        this. line 11 atm
11413 
11414~digmev Hmmm, im only on 1.3.1 at the moment so im jumping ahead with
11415        you guys (https://urbit.org/docs/tutorials/hoon/conditionals/)
11416 
11417~rabsef It's like this -> I have a string "here", I give it a face
11418        of start="here", I then have another part of my program that
11419        changes it =/ end (cuss start) into "HERE"
11420 
11421~somlus ok sounds good rab
11422 
11423~rabsef now, I could give "HERE" a face of start again, and there would
11424        be two faces in my subject w/ the same name
11425 
11426        but that is confusing
11427 
11428~somlus indeed it is.
11429 
11430~rabsef and hard to call the one you explicitly mean
11431 
11432        it is not against the rules to have two faces of the same name
11433        in the subject tho
11434 
11435~somlus ~digmev we begin sundays at 9am PST, all are welcome to join or
11436        lurk
11437 
11438        I'm surprised that it is not against the rules, but let's not
11439        over-complicate for now.
11440 
11441~rabsef in any event, that's basically why the gate gives the input a
11442        different name - you'll note that key gets the same face again
11443 
11444~somlus yes
11445 
11446~rabsef ok cool - so what does shift do?
11447 
11448        It takes in our arguments, a tape and an @ud and then
11449        immediately calls something else
11450 
11451        (this is where this program gets interesting/tricky)
11452 
11453        (operate message key (encoder key))
11454 
11455~somlus it awakens the "operate" arm to act on message and key, which
11456        requires encoder key to get awakened/activated/completed first
11457 
11458~rabsef We don't know what (encoder key) does and that means we only
11459        know that operate takes (operate tape @ud ??)
11460 
11461~digmev Im currently on the Hoon 101 and working my way through it, I
11462        will put this on my calender so i know when you guys are on
11463 
11464~rabsef let's try to figure out what the ?? is in terms of a type - how
11465        could we do that?
11466 
11467~somlus ? ~digmev
11468 
11469~rabsef To restate, operate requires 3 arguments, a tape, an @ud and a
11470        ??
11471 
11472~somlus ok looking
11473 
11474~rabsef what is the ?? in terms of it's type (like tape or cord or @ud
11475        or @tas or.. maybe something we haven't used before)
11476 
11477        nice to have all the lurkers incidentally - feel free to shout
11478        out an answer :)
11479 
11480~somlus (decoder key) seems to be a cell, but this will take a bit as I
11481        look at what decoder is
11482 
11483~rabsef (encoder key) is not a cell, it is a call to an arm w/ an
11484        argument that will eventually resolve to a value of a type
11485 
11486        we luckily have an arm called encoder - but we shouldn't start
11487        looking there
11488 
11489        let's look @ operate
11490 
11491        what does the gate take?
11492 
11493~somlus yeah this goes multuiple layers deep, to rott which uses the
11494        alpha string and key
11495 
11496~rabsef (operate tape @ud ??)
11497 
11498        operate takes a /new/ type of structure that we haven't seen
11499        before. It's called a map
11500 
11501        operate requires three arguments - a tape, an @ud and a "map"
11502 
11503        see here: |= [message=tape key=@ud encoder=(map @t @t)]
11504 
11505~somlus ok headed back up to operate. FYI I'm about 30% behind your
11506        pacing, studying the code
11507 
11508        30-50%
11509 
11510~rabsef sorry - I'll back it off a bit
11511 
11512~somlus no worries. got a bit derailed - working on it.
11513 
11514~rabsef Let me explain what we're trying to do - in ++ shift we see a
11515        call to a function that we don't know, we also see that it takes
11516        3 arguments, we know that the third of those is resolved by YET
11517        ANOTHER arm, and we want to know what kind of type that YET
11518        ANOTHER arm should produce. So, we look at operate to see what
11519        the gate takes, and it tells us that the third argument will be
11520        a map. Ergo, (encoder key) will produce a map
11521 
11522        the reason that this is important is that it will help us
11523        understand what each arm produces, in turn, and how they work
11524        together
11525 
11526~ronlyr in the docs, it says map is a Mold generator. What is mold?
11527 
11528~somlus ok map is new to me.
11529 
11530        good question
11531 
11532~rabsef per the documentation: A mold is a function that coerces a noun
11533        to a type or crashes.
11534 
11535        What does that mean?
11536 
11537~ronlyr it uses `` to cast?
11538 
11539~rabsef mold is a very good name for it. It says I have some data, I
11540        want it in this shape - stuff it in there and if it fits, leave
11541        it, and if it doesn't give up
11542 
11543        it's very much like using `` to cast but I'm not sure they're
11544        synonymous
11545 
11546        it shouldn't matter too much though, for now
11547 
11548~somlus ok so a piece of information that is a 1 but in a TEXT aura can
11549        be molded to @ud and then it changes its underlying atom-truth
11550        to become 1?
11551 
11552~rabsef Does that help at all?
11553 
11554        That is true but a mold does more than type changing. It creates
11555        structures like map which we'll discuss in one second
11556 
11557        but like you say:
11558 
11559        `@ud`'1'
11560 
11561        I guess I should try and better differentiate those in my own
11562        mind, too.
11563 
11564        I don't have a great analogy rn
11565 
11566~ronlyr so line 16 is coercing the result of encoder and key into @t
11567 
11568~somlus exactly '1' has the underlying atom of 49 because it is text not
11569        a number
11570 
11571~digmev ( I have a lot of terminology reading to do ^^)
11572 
11573~somlus so can you mold '1' to @ud and make it 1?
11574 
11575~rabsef ~ronlyr - that is basically correct only you might say it is
11576        /expecting/ a map or else it will crash
11577 
11578~somlus yeah it can be daunting ~digmev
11579 
11580~rabsef ~somlus - you can parse the '1' into @ud1 but not by casting
11581 
11582~digmev is there a faster way to tag users?
11583 
11584~rabsef ~digmev - tag them?
11585 
11586~digmev Yes like that
11587 
11588~somlus there is no autocomplete yet, they know of this requested
11589        feature.
11590 
11591~digmev Ok - though there was a shortcut nevermind
11592 
11593~rabsef ~digmev oh you mean write their name - not that I'm aware of
11594 
11595        yah no I'm just typing it lol
11596 
11597~somlus we commonly use the first 3 letters but then there is no sound
11598        on mention
11599 
11600        sup rab?
11601 
11602~rabsef ok - back to the task at hand - operate requires a map of @t @t
11603        as it's third argument
11604 
11605~somlus ok getting back to it, we are in line 16, looking at the encoder
11606        arm mention
11607 
11608~rabsef so encoder should produce for us a map, after it's activities
11609 
11610        let's talk about what a map is
11611 
11612~somlus perfect thx
11613 
11614~rabsef a map is a series of "values" and "keys" - I think of this as a
11615        vertical list of two columns
11616 
11617        on the left, perhaps, you have the value, on the right, you have
11618        the key
11619 
11620        a map will let us, from that, say "what is the value to key x"
11621        or "what is the key to value y"
11622 
11623~minder like a matrix?
11624 
11625~rabsef yes - or a two column database
11626 
11627~somlus ok
11628 
11629~rabsef so - let's conceptualize this program and see what a map might
11630        do for us
11631 
11632        We're trying to take abc...z and convert our msg to using
11633        bcd...a, for instance
11634 
11635        so "abc" becomes "bcd" and so on
11636 
11637        obviously our map is going to line up our incoming letters to
11638        their new 'shifted' letters
11639 
11640~somlus yes
11641 
11642~rabsef and then we'll parse our incoming msg letter by letter against
11643        this map to make a new outgoing msg
11644 
11645~somlus cool, I'm definitely following
11646 
11647~rabsef so - operate we know, is going to produce what (look @ cast)
11648 
11649        kethep
11650 
11651~ronlyr tape
11652 
11653~somlus yes
11654 
11655        cool
11656 
11657~rabsef bingo
11658 
11659        so we know now what operate does - operate is going to give us
11660        back our 'shifted' message abc -> bcd
11661 
11662        Here's the fun part - the cell at the beginning both shifts and
11663        unshifts your input message so if I put in abc, I get back [bcd
11664        zab]
11665 
11666        why zab? Well that's assuming maybe you put in an encrypted
11667        string.. Thus, if I put in bcd, I get back [cde abc]
11668 
11669~somlus which cell at the beginning?
11670 
11671~rabsef line 5-6
11672 
11673        sorry
11674 
11675~somlus yes
11676 
11677~rabsef the point is, shift and unshift are going to have parallel but
11678        opposite behaviors all the way down
11679 
11680        so once we follow one through, we'll be able to think through
11681        the other one way more easily
11682 
11683~somlus yes
11684 
11685~rabsef this cuts our learning of the core (7-n) down significantly
11686 
11687~somlus yes
11688 
11689~rabsef I'm trying to think what I want to do next.
11690 
11691~somlus all good so far, looking forward to understanding how the map is
11692        built
11693 
11694~rabsef one second
11695 
11696~somlus we have to define this map super precisely because it will not
11697        exist anywhere visually, i would imagine
11698 
11699~rabsef that's right - and for me I have to visualize it
11700 
11701        the two columns does the best so far
11702 
11703        but there may be other ways I haven't conceived of for maps to
11704        exist
11705 
11706~somlus I think I'm good for another 15 mins, I can feel my neurons
11707        reaching maximum "new info" capacity
11708 
11709~rabsef ok - so let's leave operate for now - we can just drill all the
11710        way to the bottom and start building back up
11711 
11712        let's go to encoder
11713 
11714        We have, from operate, passed key as the @ud that the encoder
11715        takes in it's gate
11716 
11717~somlus or I can power through, ok either way, just diminishing returns
11718        after 15 mins
11719 
11720        going to encoder
11721 
11722~rabsef we'll only go those 15 then I think we've made a lot of progress
11723 
11724~somlus true.
11725 
11726~rabsef tisfas =/ "combines a named noun w/ the subject, possibly w/
11727        type annotation"
11728 
11729        we've seen this before - here we're creating "keytape" which is
11730        a tape and has a value of (rott alpha key)
11731 
11732        we've seen the (arm arg arg) notation so we know rott is going
11733        to be an arm - further down the rabbit hole
11734 
11735        let's just go to rott and see what it's doing - we can start
11736        building back up thereafter
11737 
11738        First - what does rott want at it's gate?
11739 
11740~ronlyr tape and unsigned decimal
11741 
11742~rabsef yep - ok - rewind a second to our tisfas we just looked at (I'll
11743        type it)
11744 
11745        =/ keytape=tape (rott alpha key)
11746 
11747~somlus yes, and i'd like to input that I find it interesting that you
11748        only have to say "tape" and it knows over and over again that
11749        tape means the tape that the user inputted. very interesting.
11750 
11751~rabsef no no - aha this is where you're wrong
11752 
11753        keytape is not msg
11754 
11755~ronlyr i think tape is the type
11756 
11757~somlus oh really? ok
11758 
11759~rabsef tape is the type, correct
11760 
11761        keytape is "of type tape" and it's value is (rott alpha key)
11762 
11763        what does that first argument (arm arg arg) mean?
11764 
11765        alpha - what is alpha
11766 
11767~ronlyr keytape is added to the subject by the =/ rune?
11768 
11769~rabsef correct
11770 
11771~somlus alpha is the leg of data
11772 
11773~rabsef with a value and a potential type annotation; compare to =|
11774        which adds a named noun to the subject but does not provide it a
11775        starting value
11776 
11777~somlus the full alphabet in lower case
11778 
11779~rabsef yes - so keytape must be something about the alphabet after rott
11780        does work on it
11781 
11782        so it is not our msg, it is the full alphabet manipulated
11783        somehow
11784 
11785        let's go back to rott and finish there with our remaining time.
11786        We want to figure out what rott is doing to the alphabet, so we
11787        can know what keytape will look like
11788 
11789~somlus ok so taking a small step back |= [m=tape n=@ud] is defining an
11790        output, a cell which contains a tape and an unsigned integer?
11791 
11792~ronlyr why is it called rott?
11793 
11794~rabsef ~somlus - no - remember that |= calls a gate and requests input,
11795        it gives that input faces and states their type
11796 
11797~somlus oh yeah, I need flashcards man. I'm just going to make those
11798        because I always "best guess" these runes which is not good
11799        enough.
11800 
11801~rabsef so |= [m=tape n=@ud] says "I call a gate, or function, and I
11802        want 2 inputs, a tape and an @ud, and I shall call them m and n
11803        respectively"
11804 
11805~somlus I'll make that study tool soon
11806 
11807~rabsef ~ronlyr - I think the joke is "rotation" because that's what
11808        it's going to do
11809 
11810        but I'm not entirely sure... - anyway you could call it "flarf"
11811        and it would work as well
11812 
11813~ronlyr ok. lol.
11814 
11815~rabsef Now I've given away the game - but lets look @ rott anyway
11816 
11817        first line of rott calls the gate and requests the input
11818 
11819        we know we will be feeding it alpha ("abc...z") and our key
11820        value that we initially put into the program (let's say it's 1
11821        for now)
11822 
11823        what does line 2 of rott do
11824 
11825        (lent) establishes the number of items in a list.
11826 
11827        a tape is a list, incidentally, if that doesn't follow
11828 
11829        {a list of cords - we can review this - it gets complex}
11830 
11831~ronlyr adds a face called length to the subject, witht a value set to
11832        the length of alpha
11833 
11834~somlus sorry otf briefly answering a paitent question...
11835 
11836~rabsef bingo
11837 
11838~somlus will follow
11839 
11840~rabsef you're fine - ronlyr, ur up
11841 
11842        line 3 of rott is totally new
11843 
11844        let's look at what trim does first
11845 
11846        Tape split Split first a characters off ++tape b.
11847 
11848        Produces A cell of ++tapes, p and q.
11849 
11850        Accepts a is an atom. b is a ++tape.
11851 
11852~somlus oh cool!
11853 
11854~rabsef direct from the documentation. trim is going to require of us an
11855        atom and a tape, and produce for us a cell [p=tape q=tape] where
11856        we have split the first characters off of our incoming tape by a
11857        position
11858 
11859        let's determine what values we're sending to trim
11860 
11861        (trim >(mod n length)< >m<)
11862 
11863        assuming our input was [alpha 1], what values do the arguments
11864        of trim resolve to
11865 
11866        input to rott*
11867 
11868        (mod 1 26)
11869 
11870        modular arithmetic is weird to me still so I often try it in
11871        dojo before trying to figure it out
11872 
11873~somlus so lent m resolves to 26 which is the value then for length
11874 
11875~rabsef yes
11876 
11877        so trim now has (trim 1 >m<)
11878 
11879        and what is m again?
11880 
11881~somlus and with a key of 1 then you are chopping off the a or the z
11882        from the string. just a moment
11883 
11884        m is going to be a tape, but is not yet determined, I think
11885 
11886        it is part of the output of rott
11887 
11888~rabsef it came from our gate
11889 
11890        m is declared in the input for our gate
11891 
11892~ronlyr m is alpha
11893 
11894~rabsef and we sent in, in my example, [alpha 1]
11895 
11896        correct so m is our alphabet "abc...z"
11897 
11898        (trim 1 "abc...z")
11899 
11900~somlus oh from rott alpha key ! from line 23
11901 
11902~rabsef which will result in -> [p="a" q="bcd...z"]
11903 
11904        that is what trim gives us - a cell where the tape is split by
11905        our atom, and the new tape-lettes have faces p and q
11906 
11907~somlus that is where you are sending stuff to rott, and then rott says
11908        ok based on this, I am going to assign m to the alpha and n to
11909        the key which ia "abc...z" and 1
11910 
11911~rabsef this is helpful so we can extract one or the other part
11912 
11913        that is correct
11914 
11915        ok so now we know that s=[p="a" q="bcd...z"]
11916 
11917        and =+ "pins a value to the head of the subject" - even now this
11918        rune is hard for me to fully conceptualize
11919 
11920~somlus and then you weld them together and your new tape is bcd....za
11921 
11922~rabsef it is somehow different from =/ but it adds a face and a value
11923        to the subject
11924 
11925        correct ~somlus
11926 
11927        so weld stitches two tapes together - if anyone didn't know
11928 
11929        and we have the cell face s=>[p="a" q="bcd...z"]
11930 
11931        and we are welding q of s in front of p of s
11932 
11933        or, more appropriately we're welding p.s to the back of q.s
11934 
11935~somlus can you show me where p becomes defined as a? I don't see that
11936 
11937~rabsef that is what trim does
11938 
11939~somlus I see where s gets trimmed
11940 
11941~rabsef (trim 1 "abcde")
11942 
11943~somlus how nifty,
11944 
11945~rabsef It creates a cell w/ it's own faces interior to the cell
11946 
11947        yah it's dank
11948 
11949~somlus you get a p automatically and the rest
11950 
11951        fantastic, a great tool
11952 
11953~rabsef a lot of the standard library does that kind of stuff and it's
11954        super helpful cuz then u can address parts of the cell w/o +2
11955        notation
11956 
11957~somlus that's so great. i was trying to figure out where the "other"
11958        trim was to preserve the chopped off letter, but it was
11959        preserved automatically
11960 
11961        ok holmes, I think that might be it for now, but I want to make
11962        sure we got to a suitable stopping point. your call
11963 
11964        we have not yet created the map, but we have the regular alpha
11965        and now the shifted tape that is the new alpha-ish thing we will
11966        map to it
11967 
11968~rabsef ok - so now we're to the very bottom of the thing - we have
11969        one more thing to deal with but we know that we're going to 1)
11970        shift, which is going to 2) operate on our message based on 3)
11971        encoder which 4) is going to create a 5) rott-ed alpha
11972 
11973        sorry - yes - I was looking @ the program to try and write the
11974        above - I think that's a good stopping point for now
11975 
11976        for sure
11977 
11978        we know exactly what's going to happen from here - just as
11979        you've said - we're going to make maps of alpha and rott's
11980        product and then do work from there to shift and unshift our
11981        input msg
11982 
11983        It took me several days to figure that bad boy out the first
11984        time
11985 
11986        lol
11987 
11988~somlus cool. One critical takeaway for the non-coder or early coder is
11989        that every time we see an arm, we need to mentally attack it
11990        like it was something typed into the dojo. For (rott alpha key)
11991        we need to say "Ok we're gonna run rott now, and it will run on
11992        alpha and key, so let's make sure we understand that before we
11993        go further"
11994 
11995~rabsef yes - I think it's really important to at least get the type out
11996        of each arm's product so you can see what type the (arm arg arg
11997        arg) args are
11998 
11999~somlus dude this has been the best one yet, partly because I jammed
12000        through 12 hours of gelling the stuff from before. I need to
12001        review this today later to make sure I understand the flow. Good
12002        stuff. we can pick up next week
12003 
12004~rabsef it will help you build a mental picture of what each arm does
12005        and what it needs to do it
12006 
12007        yah - next week we'll talk about zipper hehe
12008 
12009        and coder
12010 
12011        anyone else want to ask anything before I step away for a min?
12012 
12013        I'm lying - I'm going to work on my own project next lol
12014 
12015        also - thank you ~somlus. This one is really fun because there's
12016        so much going on
12017 
12018~somlus thank you very much. I need to show the wife some attention
12019        today for being hyperfocused yesterday. will only be connected
12020        intermittently. Thank you again, good stuff.
12021 
12022~rabsef and it only gets 'worse' from here - so this is like jumping in
12023        deeply
12024 
12025        we must learn to swim, and we're starting to :)
12026 
12027        have a good day, ~som
12028 
12029~ribnys thanks for doing this "out loud"- really helps to fill in gaps
12030        in knowledge
12031 
12032~somlus thanks y'all. see you next sunday, and certainly before that
12033        here and there on mars
12034 
12035~ronlyr thanks ~somlus
12036 
12037~rabsef hey I'm glad it helps ~ribnys - that's aweosme
12038 
12039~somlus ? ~ronlyr
12040 
12041~risruc missed most of the session today, but caught what I could by
12042        scroll back.
12043 
12044        good stuff teach o7
12045 
12046~rabsef thanks :)
12047 
12048~ronlyr question about declaring variables with tislus... If I type =+
12049        b=2 (add b 2) into the dojo, is the face b being added to the
12050        subject only for the execution of the second hoon expression?
12051 
12052        when I type . at the dojo prompt to view the dojo subject,
12053        I don't see the face b, so I'm assuming tislus just makes
12054        temporary variables that go out of scope after it's done
12055        running?
12056 
12057~rabsef I think it's the way you're writing that expression
12058 
12059        > =+ s=(trim 1 "abcde") s [p="a" q="bcde"]
12060 
12061        =+ b=2 (add b 2) is not a valid use of =+ because =+ gives
12062        a face to the expression after itself for the evaluation of
12063        whatever follows
12064 
12065        my messages are lagging here
12066 
12067        There we go
12068 
12069        =+ b=2
12070 
12071        that's valid and pins b to the subject which you could call
12072        later (add b 2) and get 4
12073 
12074        oh it didn't like what I wrote - =+ b=2 is valid
12075 
12076~ronlyr yeah, laggy messages. getting them 4 minutes after the timestamp
12077        on them.
12078 
12079~rabsef does that answer ur question tho?
12080 
12081~ronlyr are you sure that is invalid? =+ b=2 (add b 2) works when I type
12082        it into dojo...
12083 
12084~somlus that would be my ship, have tried to troubleshoot extensively,
12085        one day will breach. Perhaps when OS1 comes out will need to
12086        breach.
12087 
12088~ronlyr my understanding of tislus is the result of the first expression
12089        is added to the subject so it can be used when the second
12090        expression is evaluated.
12091 
12092~rabsef I'm actually wrong
12093 
12094        yes it's pinned to the head of the subject for the evaluation of
12095        the next part and then ephemerally disappears
12096 
12097        so you should be able to, on two lines in dojo, line 1: =+ b=2,
12098        line 2: (add b 2) and get 4 and then if you b again it will
12099        find.b error
12100 
12101~bannum faces once they get used by the rune following a tislus
12102 
12103~rabsef yes lol
12104 
12105~ronlyr I finally just got that the tislus on line 45 of caesar cipher
12106        is actually the 3rd argument of the tisfas rune on line 44.
12107 
12108~rabsef I don't think that's right - so =/ length=@ud (lent m) is a
12109        complete thought
12110 
12111        it creates the length of m in @ud w/ a face of length
12112 
12113        then =+ creates an ephemeral s of [p="" q=""] for the evaluation
12114        of weld
12115 
12116        =+ uses =/'s value (length) but is not actually a child of =/
12117 
12118        does that make sense? I'm open to hearing contradiction there -
12119        lmk if I'm wrong
12120 
12121~ronlyr =/ length=@ud (lent m) isn't a complete thought. tisfas takes 3
12122        parameters.
12123 
12124        the 3rd parameter is a hoon expression
12125 
12126~rabsef it has 3
12127 
12128        =/ "length"="@ud" "(lent m"
12129 
12130~ronlyr length=@ud is one argument
12131 
12132        it's declaring a variable with the name length, with the type
12133        @ud
12134 
12135~rabsef https://urbit.org/docs/reference/hoon-expressions/rune/tis/#tisf
12136        as
12137 
12138        look in the example here - is the second tisfas the child of the
12139        first?
12140 
12141        surely not
12142 
12143~ronlyr it actually is. try it in dojo
12144 
12145~rabsef =foo =/ a 2 ; a is a valid construction
12146 
12147        I suppose you can say that everything after tisfas is a child of
12148        it as it's what is evaluated against the tisfas
12149 
12150        Is that what you're saying?
12151 
12152        ; denoting a return
12153 
12154        just put this in dojo>=/ a 2 ; then a and you will get 2. a is
12155        not a child of =/ because your > doesn't go to < indicating it
12156        needs more children
12157 
12158~ronlyr yeah, kind of. lines 45 and 46 are the hoon expression that is
12159        the 3rd argument of the tisfas
12160 
12161~rabsef but I don't think that's quite right because =/ a 2 is _valid_
12162        and closed
12163 
12164        in dojo
12165 
12166        https://hastebin.com/uwutogawug.shell
12167 
12168        you can =/ b 3 and then do any number of other computations and
12169        b will still be in the subject
12170 
12171        as long as you don't "=b" as part of those subsequent
12172        computations (which would clear b)
12173 
12174        contrast w/ =+ where as soon as evaluated, it disappears
12175 
12176~ronlyr if I do =/ a 2 in dojo, it wants another input
12177 
12178        if i type a, i get 2 back
12179 
12180        type a again, and get ford ride failed
12181 
12182~rabsef I'm having trouble following you to "it wants another input"
12183 
12184~ronlyr so, =/ a 2 is NOT valid closed in dojo
12185 
12186~rabsef you know what - you're totally right
12187 
12188        Sorry
12189 
12190~ronlyr dojo prompt changes from dojo> to dojo<
12191 
12192~rabsef yep - I was thinking of =b and I'm wrong
12193 
12194~ronlyr I just spent the past couple hours diving into understanding the
12195        difference between tislus and tisfas
12196 
12197~rabsef Well clearly I should do that
12198 
12199        Thanks for correcting me
12200 
12201~ronlyr thanks for inspiring me to dive into it!
12202 
12203~rabsef :) sorry for bad information, truly. I should be more careful
12204 
12205~sibhul Is there an optimal place to look in the hoon tutorials to look
12206        for an example or explanation of using a naked generator to
12207        apply an aura to an atom?
12208 
12209~rabsef https://urbit.org/docs/tutorials/hoon/nouns/
12210 
12211        I would look @ the casting part there
12212 
12213        and this example of a naked generator is overly robust for your
12214        purposes - you don't need recursion - but you could get the
12215        style/format her:
12216 
12217        here*:
12218 
12219        https://urbit.org/docs/tutorials/hoon/list-of-numbers/
12220 
12221        What part are you feeling you're having trouble w/?
12222 
12223~sibhul I'm having trouble with the aura return in a *.hoon. I'm fairly
12224        certain that I have the input worked out, but the syntax of aura
12225        return seems to be elusive.
12226 
12227~rabsef wanna pastebin your code and I'll see if I can give a hint w/o
12228        giving it away?
12229 
12230~sibhul Sure. I've already got my embarrassment prepared.
12231        https://pastebin.com/fiKRhgTj
12232 
12233~rabsef I'd review how casting happens in the nouns link above
12234 
12235        You're pretty close. Also you might consider how the list of
12236        numbers syntax varies from your long form into tall form, which
12237        I think is easier to read in this case
12238 
12239        last hint, I just got your code to work by changing some things
12240        and mine only has one set of parentheses
12241 
12242        (it's still in long form)
12243 
12244~sibhul Yeah, I can feel I'm close, something isn't quite clicking yet.
12245        Going to take a quick break and come back to it.
12246 
12247        Okay, I think this must be darn close. It now 'runs' with no
12248        syntax errors, but I get 'ford: %slim failed:' as well as other
12249        errors. https://pastebin.com/V2j3NpZn
12250 
12251~rabsef you're very close but you've sort of half changed to tall form
12252 
12253        tall form should be like |= required input <return> things to do
12254 
12255        another way of saying the above is you removed different
12256        parentheses than I did
12257 
12258        https://pastebin.com/HxYdd1bw
12259 
12260        ^I didn't fix it or help you - just showed the different forms
12261 
12262~sibhul Got it! Thanks so much. Not sure why that didn't register after
12263        reading the reading the style guide. Much appreciated.
12264 
12265~rabsef I think tall/long/irregular forms are tough, myself
12266 
12267        glad you got it
12268 
12269        I'm gonna go do nothin for a while - see yah later hoontards
12270 
12271~ronlyr on line 41 of the Caesar Cipher, chart (~(put by chart) i.a
12272        i.b), I understand that it is assigning the head of a as the key
12273        and the head of b as the value to the map called chart, but what
12274        exactly is the ~() syntax doing? The only reference I can find
12275        to ~ is that it means a null value.
12276 
12277~bannum ~(put by chart) is shorthand for climbing down the structure, so
12278        it'll become chart.by.put or something
12279 
12280        I forget which direction it is its been a while sorry
12281 
12282~hastuc ~(put by chart) is syntax for accessing the arm in a door
12283        (https://urbit.org/docs/glossary/door/)
12284 
12285        ~(put by chart) gets the door named by, sets the sample of the
12286        door to chart, then gets the arm 'put' from the door
12287 
12288~rabsef check out 2i for got by in use
12289 
12290        its irregular censig I’m pretty sure but I can’t exactly explain
12291        it. that’s one I need to understand better
12292 
12293        welp I was way behind there lol
12294 
12295        hastuc correct me if I’m wrong but the gate created by the arm
12296        put of the door by then needs an argu so you get (the created
12297        gate arg I.a I.b)
12298 
12299~hastuc yeah that's right
12300 
12301~rabsef I wanted to follow up on this cuz I was looking @ the by arm of
12302        the standard library which is a wet core and I was like "wait
12303        but that's not a door" - long and the short of it is a wet core
12304        is a door
12305 
12306        so just don't be confused by by not having |_ in it
12307 
12308        in ~(got by map) the map replaces the a value of the by arm's
12309        sample. There's more to that but that's what I wanted to
12310        establish for myself - IDK if that' ssuper helpful, but maybe so
12311        :)
12312 
12313~walden does anyone have an estimate of the number of people who know
12314        hoon? just curious
12315 
12316~hastuc well 31 people have committed more than 1k LoC to
12317        urbit, so like at least 31 intermediate-advanced
12318        hoon developers with sufficient time to contribute
12319        (https://github.com/urbit/urbit/graphs/contributors?from=2013-09
12320        -22&to=2020-02-11&type=a)
12321 
12322~walden (sorry for the u-h crosspost, ignore)
12323 
12324~norsyr ha! I made the cut! (barely...) I'll add intermediate-advanced
12325        hoon developer to my linkedin profile :D
12326 
12327~somlus very cool to know that actually ~hastuc
12328 
12329~risruc +1