Hi @Adám. Quick FYI: Safari somewhat jumbles the APLcart link and "query:" text.

(Resizing the browser window does not fix it.)

@TorstenGrust Thanks. I'll try to do something about that when I get my hands on an Apple device.

@TorstenGrust better now?

@Adám Looks good on my iPhone (SE, small screen, latest Safari available)

just happened upon a Guy Steele talk from 2011 about parallelism and hist Fortress language, with a comment by Torsten Grust ; it's a small internet :)

I half remember asking here a while ago, about a way to group things into N pieces in APL; abcdefghi to abc def ghi and that there was a surprisingly simple answer, like an N-wise reduce

but I can't come up with one, and then I found codegolf.stackexchange.com/a/180987/571 and if that's it, I think it's a poor memory

@J.Sallé Interestingly, TorstenGrust's screenshot looks quite different from what I see in FFQ/Edge/Chrome on Windows — they all render more or less identically. Edge has slight alignment problems for the × and the contrast, but that's it. Torsten is seeing large distance from X, Y, X down to the results, an then of course the cramped query.

@TessellatingHeckler There is {⊂⍵}⌺(⍪3 3):

⎕←{⊂⍵}⌺(⍪3 3)⊢'abcdefghi'

@Adám
┌───┬───┬───┐
│ ab│cde│fgh│
└───┴───┴───┘

But as you can see, it begins to far left.

@Adám that appears to rely on knowing in advance how many chunks there will be

Stencil, aha

@TessellatingHeckler No, ⍪3 3 means window size 3, step size 3.

oh ok

I still see Stencil as witchcraft.

@TessellatingHeckler dzaima/APL and Extended Dyalog allow a placeholder value in the left argument to ⍴ to auto-size that dimension.

@Adám what is "Extended Dyalog", exactly?

is it things you wish were in the normal one, but haven't made it in (yet)?

That's a very good summary of it, yes

@TessellatingHeckler Me doing fancy regex-fu to insert my covers that extend the domain of Dyalog primitives, and adds a few new ones.

or things which are in there, but aren't APL-2 / backwards compatible?

@TessellatingHeckler APL2 hasn't really added much, so there's basically no compatibility issue.

@Adám neat; regex on the input source code, like a pre-processor?

@TessellatingHeckler Very much so. The code is here

#tio do apl-dyalog-extended ⎕←↓¯1 3⍴'abcdefghi'

@Adám  abc  def  ghi

from ⎕R to⊢code should be in the header of Extended's repository as a TL;DR

@user565251 Hi preaburn. If you want write access to this room, email me: adam@ with the same domain as www.dyalog.com.

@J.Sallé Hehe. True, but it can do so in multiple ways: REPL, "⎕FIX", "⍎".

Well yeah, but none of those capture the essence of the regex-fu going on in there

@TessellatingHeckler A good example:

@Adám TIL ⍧

@dzaima Like it?

@Adám seems useful

@dzaima Also monadic ∩. I couldn't decide which monadic function to make that, nub-sieve or self-classify, so I added both. Nub-sieve inspired me to count-in, and they fit well together, so they got ⍧. The fact that it looks like an overlaid C and I is just happy coincidence. I anyway wanted a shoe-y symbol, and didn't need a pair (there is no ⊃| combo.)

if i understand monadic ⍧ correctly, it's "is this item the first occurrence of itself"?

@dzaima Interesting way to put it. I think of ⍧Y as the mask that will give ∪Y.

Kind of like ⍋Y is the permutation that will give ∧Y

@Adám that doesn't necessarily include that ⍧2 2 2 should be 1 0 0, not e.g. 0 1 0

@dzaima True. But ⍋2 2 2 could give 2 1 3.

@Adam, yes, better now.

@Adám but it really shouldn't, just as above

@TorstenGrust It can pass. On other devices (browsers?), there is a quite sizeable gap now. Also APLcart and query appear vertically misaligned, and there's too much distance from X,Y,Z down to the results. Hm. I'll have to investigate one day, but I guess it is "good enough" for now.

@dzaima One could say that ⍧ and ⍋ are greedy.

@Adám I think they just look misaligned because of the underlined APLcart

can't be 100% sure

@J.Sallé The top of "car" is definitely higher than "query"'s top.

Yeah, but there are some fonts that put a pixel or two padding between the characters and the underline. Again, can't be 100% sure just from the image

@J.Sallé the bottom of letters should always align, underscores and character parts below the bottom shouldn't influence anything

@dzaima But they are not at all connected in the HTML/CSS.

APLcart's utility (to me) is phenomenal. Definitely deserves some stage time, if you ask me.

@Adám sure, but imo that still should be the case (though i'm no designer)

My FFQ/W10 inspector says both __APLcart __ and query have the same line height and position within the container. Might only be happening in Safari, though

Markdown let me down there, but you get the gist

@TorstenGrust Thank you. I hope to increase it's value by a lot, when I add all the FinnAPL idioms. About a third done now.

@J.Sallé my chrome also draw both at the same height

Yup, even Edge aligns them.

Chrome on macOS.

@TorstenGrust Ah, so it is only Safari on macOS.

@Adám Yes, it appears so.

Safari's rendering might be weird.

@TorstenGrust But APLcart and query are very close to each other. Maybe font rendering is different on macOS.

"might be" being the understatement of the century here

@Adám Hmm, since I'm living in an macOS-only environment right now I cannot compare...

@TorstenGrust FFQ/Win10:

@TorstenGrust It is responsive. If there's enough room, it gives less dense explanations.

@Adám I didn't notice that until now! Great. (Works on Safari, too.)

implemented ⍧ with the lazy/obvious hashmap/set solutions for general types for now, sometime i'll get around to specializing for specific types

@Adám what is the intended use / who is the intended ideal user for APLCart?

@TessellatingHeckler It is intended for both experienced APLers and those that just know basic APL syntax, to answer their How do I… questions.

@TessellatingHeckler Thanks for asking that. I've added a line to the README.