Is this pornography? bonfire.com/co-dfns-thesis-edition/…

<moon-child> is there any thematic difference between the )-prefixed metacommands and the ]-prefixed ones?

moon-child: I think ) commands are supposed to be built into the system while ] commands are configurable by the user. I don't get what you mean by "thematic".

<moon-child> @Bubbler ah, ok, that makes sense (and explains why using nonexisting ]-commands spits an error about ⎕se.ucmd)

For Dyalog APL, ] commands are indeed the wrapper for functions pre-loaded in ⎕SE.UCMD.

Other APL implementations might have their own way of handling them or have no support at all.

this answer is interesting

Which part is interesting to you?

it uses a while loop

is indentation required in apl?

I don't think so

And almost any kind of while loop can be golfed by using dfns with recursion or ⍣ instead.

yeah that's the point

it uses a "traditional" apporach

pretty rare to find stuff like this

Yeah, mainly because it's way too ungolfy

and the site is Code Golf

I guess it could as well be deleted for not being a serious contender.

Also there's this

Meanwhile, as this is still on hold for the bounty claims...

CMC: Explain the 14-byte answer in this link.

@TessellatingHeckler APL#?

@Adám that's what was in my mind thinking of it; the wiki page says that was a bit of a departure and compatibility break from traditional APL so that probably does count for what I was thinking of

Switching to pass-by-reference as default is definitely a huge departure.

@Bubbler wait thats my answer smh

¯_¯

@Adám curious, did regular (as opposed to modified or indexing) assignment modify the reference, or rebind in the local scope?

@Moonchild Modify the reference, I think. Have a look in the linked docs there.

⋄⍎¨8 5⍴'6547232640161280812803824015960052700003'

@Razetime RANK ERROR

⋄{⍎¨8 5⍴⍵}'6547232640161280812803824015960052700003'

@Razetime RANK ERROR

fine then

tio.run/##SyzI0U2pTMzJT/…

How do I make it execute each row instead of each character?

@Razetime split the matrix into its rows, and execute each of those

(or execute ⍎ for each cell, but above is probably shorter)

@Razetime Rows have rank 1 (they are vectors), so ⍎⍤1 ― or ⍎¨↓

nice

found a better way to make the matrix

how do I include unprintables in an APL answer?

@Razetime You can use it as-is if they're part of ⎕AV, otherwise you simply can't

@Bubbler What do you mean?

@Razetime Inside a string you can use pretty much any character.

Well, more precisely, you can't use 1-byte-per-char scoring

(and also you can't have newlines inside a string)

@Bubbler Yes you can, if you can figure out a way to put them there.

Wat.

@Bubbler Try it online!

⋄⎕AV[128 192 224 240 248 252 254 240 216 152 12 12 6 6 3 3]

@Razetime Illegal code

last two chars are unprintables

How do I include them in a string

Should I get an xxd?

No, that's fine. The last two are CRs.

oh

my bad

@Razetime Try it online! (Don't do this in production code…)

is there no correct way to just include the carriage returns as single chars?

@Razetime They are just single characters, and this is valid code. Your problem is the editor, not the interpreter.

I'm just using a roundabout way to coerce the characters into the string in the function. After that, it is a normal function with a normal string that contains literal CRs.

I don't think the interpreter will accept the code when given as a plain source code file, which is a requirement for code golf scoring

@Bubbler I don't think that's correct. When we use Dyalog Classic, we go by how the interpreter represents the code, not how it is in a text file (which isn't its native format anyway).

Here's what I came up with for now

@suffianfarooq Welcome. Interested in APL?

@Razetime That's good. Note (⊥⍣¯1)→⊥⍣¯1⊢ or use Extended's monadic ⊤ and remove the 2 too.

@Razetime You can "unswap" ,⍨ since the left arg needs parens anyway.

Ok, I think I can post this.

Should remember extended exists lol

what i was thinking is: since the last 4 chars are in ⎕AV, they should be countable as 2 bytes instead of the workaround shown.

I'll count it as 36 bytes for now.

You can't use ɫ with SBCS

not in the codepage?

I guess I've got something smaller though it exclusively works in Extended

tthat's what I tried

but encoding in binary was easier to figure out compared to that

@Bubbler very close to 05AB1E

But then we need to map 0s and 1s to blank and some char

the question says a matrix is fine right

oh it says characters

Try it online!

now it should be fine

Yeah, that works at least

<loke> What on earth is APLBot? It seems to be piping some other channel into this Freenode channel, and it's utterly unreadable.

@Adám

seems to be a problem here?

loke: It connects to a Stack Exchange chat room (chat.stackexchange.com/rooms/52405/the-apl-orchard) ― besides for evaluating APL.

In what way is it unreadable?

<moon-child> loke: it connects to the apl orchard, on stack exchange chat.stackexchange.com/rooms/52405/the-apl-orchard. There's a bug where links aren't properly detected, but aside from that it should work fine

Let me try: Try it online!

Oh, so the link part is truncated (and only plain text is visible)

<loke> It's unreadable because all messages seems to come from APLBot. A proper link is what Matrix does where matrix users are real users on IRC.

<moon-child> ok, but neither SE chat nor IRC are matrix

@Bubbler oh, 0-indexed.

And chat message edits are not reflected either

It'd be a neat enhancement if the bot would detect edits and update APL evaluations too.

@JoshD Hey, nice to see you here.

Is that a dyalog person?

@rak1507 Yes.

Anyone have an APL TIO template?

For code golf

@rak1507 https://tio.run/##SyzI0U2pTMzJT////3/ao7YJj3r7HnU1P@pd86h3y6H1xo/aJj7qmxoc5AwkQzw8g/8DAA

does this work properly? the tio link

yes it's alright

Looks good.

@rak1507 I like your mathematical and array-oriented approach there.

Thanks! I originally had a ton of parentheses around recursive calls

Hey, I like lurking here occasionally :)

codegolf.stackexchange.com/a/211666/95516 can this be golfed at all?

@rak1507 I think you can remove () and replace with ⊢ like {⊃(1↓⊢,+/)⍣⍺⊢⍵} to save 1 byte

ah nice

It behaves the same in your TIO examples

@rak1507 You can save 2 by converting to full program: replace ⍺ and ⍵ by ⎕ and remove braces.

feels like it should be possible to make it into an atop like ⊃⍤((1↓⊢,+/)⍣⊣) but I can't

@TessellatingHeckler yeah, the right operand of ⍣ is an unfortunate case of impossible-to-tranify

@Adám ah nice, not sure how to go about TIO-ifying that, can you?

@dzaima ohhh, oh well

@rak1507 Try it online!

Thanks!

@rak1507 for a full program, you put ∇f in header and ∇ in footer

@dzaima (it's because it handles a function and array operand very differently; BQNs f⍟n is "repeat n times" always (evaluating once if its a function) so …⍟⊣ would work)

then type the function name (f) and input separated by newline

my BQN interpreter port is awash with scope issues; there's a lot of pushing lambdas onto stacks and evaluating them later, and PowerShell is dynamically scoped and all the surrounding runtime context is different. Adding some variable shadowing and closures and a couple of global variables has nerfed the most direct failures, but now i think there's things quietly being $null and changing the behaviour without erroring

@TessellatingHeckler Ouch. Guess you'll have to switch from using $null for Nothing (if you haven't already) and check for it in call?

@Marshall I haven't been explicitly using $null; it's things like let c = sv => vm(st,[sv.concat(v),e]); is defined in JS inside a block where it has access to vm the virtual machine executor function and B the bytecode array. When it comes to runtime delayed execution this c function, the vm variable is null because it doesn't exist in the runtime scope.

trying to run it throws an error. With a closure to capture vm so it can get to it, there is a function to execute, but in that vm function the bytecode array $B is blank at runtime.

It is interesting to work through; the fact that there's lots of stacks was a surprise

PS is such a mishmash language, its bolted-on class system switched to lexical scoping, at least a bit. might investigate that if I can't get much further this way

'02210313074020029000033739' not sure how good apl is with kolmogorov complexity type things but can this string be golfed at all?