Gives a higher order array, say z←(2 2 2⍴⍳8), is there a good way to turn it into a list of major cells? I.e. in the case of z, we would get (2 2⍴⍳4)(4+2 2⍴⍳4).

@B.Wilson Before I answer, can I ask you what you searched for on APLcart?

@Adám Oh. That really helpful compendium of APL knowledge that you can very conveniently search directly from within the session? I totally didn't forget about that :P

@Adám But then what would you do for the icon?!

Ꜳ

Hahaha! You're too good!

@Adám alternatively, rename the Orchard to "ask about things APLcart doesn't have"

The chagrin is real.

while I do understand that it would be a massive undertaking as well as probably unmanageable, I'd love a polyglot APLCart for multiple array languages

so I can lookup an idiom from J and see how Dyalog spells it

I have common function names from various langs as search terms. Could add J primitives

what I would love is to look up "cartesian" and see +∘(¯11∘○) with a little triangle and then when you click it it opens a drop down with "J: j." etc

also maybe I can't find the right keyword but this specific example doesn't seem to be on APLcart, neither as this nor as 0j1×

if it's there, I'd recommend adding cartesian and rectangular as keywords

@RubenVerg aplcart.info?q=join%20complex?

yep, I would have never written "join complex" for that

BQN has ⍳, right?

does Extended/Vision have a primitive?

Extended has ⌾. However, they should be ⊕ and ⊗.

@RubenVerg You mean index generator/index of, or something for complex numbers?

@Adám bqn's iota is J's j. iirc

afaik, BQN has no complex number support.

Sometimes the simplest things can be such a joy: apl(){ echo "⎕←$*" | dyalogscript /dev/stdin;}

@Adám "optional extension" which I guess no impl supports

I was considering doing BQN-style casing enforcement without role switching for tiny APL, is it too radical?

@RubenVerg No, that's a good idea. Will make parsing possible/simple.

That's what I recommend too.

what's a ref?

I'm also considering generic syntactic stuff like alpha/omega underbar for operands instead of double alpha/omega

@RubenVerg a pointer to an object.

@Adám like something from quad new?

Yes, ⎕NEW will always return a ref, but there are many other ways to get references to objects.

(none of which I will have!)

not having pass by value everywhere would massively complicate an implementation

@Adám what other ways are there to get refs?

E.g. 0⎕JSON'{}' and there are always-available refs like ⎕THIS, ⎕SE, #, ##. Then there's ⎕WC which can create objects with ref pre-assigned to a variable, oh and ⎕NS⍬ returns a generic object.

oh right I always forget namepsaces are references

too late for this year, but has having copyable tests for part 2 problems ever been considered?

also, I don't know if it's just me but there seem to be some SSL issues, happening both on my phone and computer, connected to two different networks

(cc @Adám, sorry for the maybe useless ping but too little time left before the end of the competition)

@RubenVerg What do you mean by copyable?

@Adám instead of just showing the result, say something like "it should return 2 3ρ1 2 3 9 8 0" or something

OK, I'll take it up for next year's contest.

great!

looks like each request has some random probability of succeeding with ssl or not, which seems to suggest it's a problem on my end

what's the idiom for the bounds of an array?

Bounds?

isn't the amount of elements called the bounds of an array?

idk if product of shape or tally of ravel is better

×/⍴ would be more efficient.

alright thanks

can you confirm the issues I'm having on the contest site are just on my end?

I've not experienced or heard of any issues but yours.

perfect, I'll check more thoroughly for issues on my network then

sorry for bothering :)

I belive it is running on a large data centre right off the UK–France backbone.

I expected it to run on a random Thinkpad in the Dyalog headquarters basement :p

We do have non-critical stuff running from our server room in Bramley, and they occasionally see outages (usually someone digging through the fiber). However, we're investing in a fallback connection using the mobile/cell towers.

Is there an idiom for Div by zero produce +Inf or -Inf according to float point calculation in APL?

In an APL that doesn't have infinities? Do you want max/min float?

With ⎕DIV←1 then ÷+(⌊/⍬)××⍤⊣×0=⊢ does the trick. Maybe something nicer exists.

@Adám is there some other way I can submit my sols for the contest? site still isn't working for me, i'm sure it's an issue on mine but can't seem to fix it

@RubenVerg Send them with an email to the address listed on the site, and explain your situation.

thanks

@Adám oh no. I really hope Dyalog has NaN and Inf now.

Nope. ngn/apl does do the infinities out of the box, though, but 0÷0 gives a domain error.

Other than that I also want to handle ¯3○0J¯1 (arctan), I figured J might be more suitable as it can do (-3)&o. 0j_1.

Which gives 0j__

@RabbiKaii Hi there. Interested in APL?

@Adám nothing but a curious peek. Keep up the good work. Easy fast, lashana haba birushalayim

Another time maybe. And same to you!

@RabbiKaii Did you come here from FrumDevs?

@Adám I came to this pardes from Mi Yodeya :)

Aha.

Array Partitioning isn't meant to be io-agnostic, right?

you mean ⊂⊆? they don't care about ⎕io

no, I mean challenge 2.2.4

oh, there's nothing in the spec to do with ⎕io. assume ⎕io←1 in your solution or localise it to 0.

like, if the third element of spec is 2 2, I shouldn't assume that if io is zero it wants to start from the third row and column?

> Your solutions will be tested in the default Dyalog environment using (⎕IO ⎕ML)←1. Your code can employ a different, localised, setting for either of these if necessary.

right

thanks

@Adám email sent - hope i've explained enough :)

Second draft for the first TinyAPL article - the code now actually compiles!

also added a prerequisites paragraph

I would love for the articles to be valid Literate Haskell - but having multiple modules makes that (afaik) impossible

Lol, J knows how to do _3 o. _ but not _3 o. 0j_