The APL Orchard - 2020-09-20

« first day (1334 days earlier) ← previous day next day → last day (1303 days later) »

12:16 AM

@Marshall I agree it would have been more consistent if empty string was true, but those were the rules...

@Jonah Nothing wrong with that, but the right way to specify it would be to put "non-empty" in the definition twice instead of adding such dubious clarifications. Not very important, of course.

@Marshall btw your idea still saves me a few bytes *@#*]-:]&.(,~inv), so thanks. I had no idea ,inv has that obverse, even though I tried a few related verbs which didn't have one. In general, how can I view the definition of an obverse or view a list of the builtin ones?

@Jonah Check out ,~ b. _1.

ah right, thanks

@Jonah Also, #<.]-:[:,~,~inv cuts two characters, at the cost of losing the elegance of Under.

12:32 AM

Nice.

11 hours later…

11:21 AM

⋄{(⊢⍪⌽[1]∘⌽)(⊢,⌽∘⍉)⍵↑⍨-2/⌈/⍴⍵}↑'Hello' 'World'

@Razetime 0x0.st/ila_.txt

⋄{(⊢⍪⌽[1]∘⌽)(⊢,⌽∘⍉)⍵↑⍨-2/⌈/⍴⍵}↑'APL' 'FUN'

@Razetime
   FA
APLUP
FUNNL
 LNNUF
 PULPA
 AF

I made this C4_4 reflection function thing

whats the best way to simplify the left part of it?

the transposing and stuff

@Razetime ⋄⌽[1]4 4⍴⍳16

11:33 AM

@dzaima
13 14 15 16
 9 10 11 12
 5  6  7  8
 1  2  3  4

?

@Razetime what does the above look like?

flipped?

oh it's a 4x4 table?

@Razetime no, that was the correct answer. but i'm unsure how to continue on hinting :p

ah the struggle of teaching

11:38 AM

(the answer is this; guessing you just forgot, as i'd expect bracket axis being a way later thought than that)

reverse first...

obviously theres function for that aaaaa

@Razetime as more golfing opportunity, with some rearranging, you can drop the need for ⊢

ok lemme see

thats what i was wondering

how to combine (⊢⍪⌽[1]∘⌽)(⊢,⌽∘⍉) into one

@Razetime i actually keep the two separate (each doing mostly the same thing); what's changed is the exact expression in each parentheses

so you mean, you change the order in which the rotated parts are made

11:46 AM

@dzaima (and also no ∘; both expressions are 3 chars)

bruh

@Razetime while the exact intermediate output is different, the first part still outputs an n×2n matrix, and the second, 2n×2n

@dzaima (so {(fgh)(FGH)…})

bit confuding

(comparison of your intermediate output and mine)

Hard to figure out what an ASCII art language creator would do lmao

@dzaima I'll try making something shorter first

at least a byte

11:57 AM

@Razetime it's actually a pretty common trick for things like (⊢[,⍪][⌽⍉⊖]∘[⌽⍉⊖]). It gives a different output, but that's usually fine

@dzaima Got it!

@Razetime yep

gr88

got another -1 by changing to a train and using an alternate method for getting 2/⌈/⍴⍵ (that doesn't do the same thing on rank>2, but which doesn't matter here)

lol have fun

12:12 PM

@dzaima (fwiw that method uses an ⍥ and doesn't save/lose any bytes in a dfn)

12:28 PM

A commenter points out that "scalar" doesn't match the term from linear algebra (I knew this) and suggests using "box" instead from an array of rank 0. I don't really want to use box because it differs from the usage in boxed array languages like J, and I don't think it properly conveys the idea that there can only be one thing in it. Maybe "site" would be better? "Frame" or "shell" might work but also have conflicts.

May also be nice to have a term for any value for which rank returns 0, that is, an atom or site/box. "Spot" and "spec" sort of make sense, but sound weird in sentences.

Unit. An atom or site is a unit.

@Marshall i like unit. but i don't think it's worth having 3 terms for atom/scalar combinations

(imo scalars shouldn't be that "important"; they're either used as a "rank 0 array" in which case calling it as just that is useful, or as a special-case arg (e.g. avoiding needing each-left/each-right))

12:45 PM

@dzaima So you'd just go with "atom", "unit", and "atom or unit"? My thinking is that "unit" for the last of these wouldn't be used to often, but its easily understood when it does show up.

@Marshall i'd probably go with unit including atom, and no name (besides "enclosed") for a non-atom enclosed/boxed thing. atoms and non-simple scalars usually behave the same way, the difference usually being in making them

Oh, "unit array" works just fine for a rank-0 array.

@Marshall that works too

(i like "array" being in the name a lot, as that's the important part. Miles better than "non-simple scalar", but that's a pretty unfair comparison)

1:30 PM

@dzaima the atop thing right

@Razetime "over" thing. A(f⍥g)B → (g A) f g B

much golfed (1858→1453); turns out there's a way shorter way to check for non-equality - (1↑1--↑---) while previously it took 35 chars
https://tio.run/##zVU7bsMwDN17Co8kQAJhl9yg9whQoEuAdu0FshVoz9Cb@SIpH/WxLNtpAQ9NHFgyP0@PTyJ0ejvr8/vp/PpyvT6Nl08yVWMikvHyxeRzH4XjI/9MjYwNI6swWUQIkY/jxzdyPBZWn2pBIIUvcmCG253@VlXO2KKcPHkZeDtgkzzrGTJn/EhvfTSR8HTTBi3WztVVok14oiKbfmkrlFSYxJN4hFSzSrOGCODOJVAGnjXXgmW7NZmiFSGQSplB@liHXdZ2A5Vq1hr5FZm6A/CbllsBpDTfjxkrhuxA0bpU2gM/TPvq3cANZbFNtg9@Ax9GgO/CDkvbjYtmJK3tWFokUahnVO@rE2/L9T9qIfp@BUNgqKN/bP1c0YSkfffO@4M71WIRZJRE/2Jvk@ppMGG1uETijYAI0jSzElWSZ5lhyc4Ur/M/rwBWWlMIBt@j…

1:51 PM

what in the WORLD

@Razetime a lot is possible with the abstractions

2:11 PM

Hi all I need some help with John's depth first search in APL. I am wondering why the recursively defined operator dfs←{⊃∇⍨/⌽(⊂⍺ ⍺⍺ ⍵),⍺ ⍵⍵ ⍵} would ever terminate?

His video can be found here: youtube.com/watch?v=DsZdfnlh_d0

@jimfan recursion terminates when the argument to ∇⍨/ has length 1. of course, that's at the mercy of ⍺⍺ and ⍵⍵. an incorrectly implemented ⍺⍺ or ⍵⍵ could cause non-terminating behaviour.

correction: only ⍵⍵, not ⍺⍺

2:32 PM

@ngn Thank you. I am guessing NARS2000 behaves differently in this regard. The dfs is giving "STACK FULL" when executed

@jimfan no idea about nars

@ngn Nevermind, I am checking out NARS2000 doc

2:54 PM

@dzaima is this for the turing completeness question?

@Razetime it was a reply to it

oooo

2 hours later…

4:30 PM

Is there hashing built in to APL? Couldn't find anything

hashing, meaning?

Hashing algorithms, MD5, SHA1, etc

4:59 PM

@rak1507 don't think so, aplcart doesnt seem to have anything n it

I also checked the nars docs

3 hours later…

8:11 PM

@Razetime You can run Extended locally in your IDE too.

@rak1507 There's the Dyalog Cryptographic Library.

8:59 PM

Jeez, did you use a script to generate that?

9:53 PM

@rak1507 I don't know. It does look very terse, but it isn't really so bad. I figured out how to use it with very little experimenting. I can show you my code, if you want.

github.com/Dyalog/DCL/blob/master/ASN1.dyalog the OidTab stuff there must be super slow, if it has to run that when it loads? Right? Idk

i'd be interested to know how long it takes to hash a megabyte

« first day (1334 days earlier) ← previous day next day → last day (1303 days later) »

Transcript for

Sep '2020

The APL Orchard

apl.chat ― Learn, teach, ask, code, golf, & discuss usage. See ...

apl array-manipulation j tips

join this room
about this room

00:00

06:00

12:00

18:00

all times are UTC

site design / logo © 2024 Stack Exchange Inc; legal

mobile