CMC: Beta function. E.g. 2 B 4 gives 0.05

@Adám Does Dyalog have an integration op? If so, it must be very obscure as I've never seen it

@J.Sallé Not built-in, no, but the idiom search gives A+.×⍎XV,0⍴X←NV[1]+(A←--/NV÷NS)×0,⍳NS and iirc, one of the APLPress books lists an integrate operator and a derivative operator.

@alexz02 Hi and welcome. Interested in APL?

@alexz02 No problem. Let us know if we can help with anything, and feel free to use the bot.

@TorstenGrust You do realise that those keywords are just APL functions after a very thin pre-processor, right?

@Adám could you explain to me how to make Extended's ⊤ work like ⊥⍣¯1? You sent me a piece of code a few days ago that did it, but I'm a little confused about it.

@J.Sallé 4(⊥⍣¯1)1234 in regular APL is 4⊤1234 in Extended.

@J.Sallé It is actually a mistake in Extended, but I think I'll keep it.

@Adám so I must be doing something wrong

@J.Sallé Extended only extends Header/Code/Footer. Input is good ol' APL.

Oh, I see!

@J.Sallé Try it online!

@Adám Yay for random test cases!

@J.Sallé :-) Also: '()[]{}'⊇⍨6∘⊤

Or even 6∘⊤⊇'()[]{}'⍨

Never used ⊇ before hahahah

@J.Sallé I call it "sane indexing"

It's like ⌷ that accepts vectors as left argument?

@J.Sallé Yes, it is exactly that.

Nice. I've always tried to use ⌷ for that and it always blew up in my face hahahah

@J.Sallé The trick is to enclose the left argument: (⊂indices)⌷data

Muh bytes though D:

Yeah. It can be written tacitly too, but ugh: ⌷⍨∘⊂⍨

It's very ugly hahahahaha

… and unreadable if you have not seen it before.

@Adám or as ⊂⍛⌷ with your extension mentioned before

@dzaima I did add ⍛ but I need to tell Dennis to pull. What do you think of ⍛ ?

@Adám it seems useful, and i'm working on adding it to my apl right now. The dyadic ∘ on it's own feels a bit strange

imma leave monadic unimplemented though

Yeah, but together with ⍛, ∘ seems much more acceptable. Unfortunately, they symbol ∘ was wrongly assigned. It should have been ⍩ like Iverson inteded which gives the obvious symbol for the companion. But now that we have ∘, ⍛ was the best I could find.

@dzaima I've (tentatively) made f⍛g Y be g Y but I'm not sure if it that's useful.

Hm, never though of the fact that monadic ⍸ is /⍥,∘⍳∘⍴⍨

aw, there's no <+¨ character.. there goes another option for the left/right-each chars

@dzaima Yup. Unicode included only what Iverson used.

@Adám Yes, that was obvious.

⌷⍨∘⊂⍨ is to be read as (⌷⍨∘⊂)⍨ and not (⌷⍨)∘(⊂⍨). What's the APL syntax rule that governs this?

@TorstenGrust operators bind left-to-right

So, the leftmost ⍨ binds its argument, then ∘, then the rightmost ⍨?

@TorstenGrust yep, exactly the opposite of regular functions

Aha! Thanks, @dzaima.

(except that regular functions are ambivalent but operators aren't)

@dzaima Also (as of yet) Dyalog doesn't allow implicit currying for functions e.g. 2* but does allow implicit currying for operators e.g. ⍣2

@Adám does "as of yet" mean it's being considered?

@dzaima It is.

:o

@dzaima Some are just concerned that there is a more useful meaning for A f than A∘f since the latter is already available.

@dzaima We could have defined A(f g h)B as (f A) g (h B) which kind-of is to the current definition as ⍤ is to ⍥

However, (f A) g (h B) can be written as f⍛g∘h but how would you write (A f B) g (A h B) ?

@dzaima ACK.

@Adám do you have examples of what else could e.g. (2+) mean, as anything i think of feels strange/unintuitive? but adding that does seem to hint at f←{a b c ⍵} equaling f←a b c, which could be considered as bad

@dzaima added, along with a fancy live grapher because i had an idea, and just could not not

In Haskell, (2+) is λx.(2+x).

@dzaima We should invite you to speak at Dyalog '19!

@TorstenGrust Does 2.+ mean anything?

@Adám Nope, that's a syntax error (because of the trailing . after the literal 2).

@dzaima Just because I can't think of anything, doesn't mean there isn't another option… But consider 2 f/Y which is f/¨2⊂/Y so maybe removing the reduction from 2 f/ makes 2 f such that (2 f)Y is f¨2⊂/Y ?

2 . (+) does have a meaning, but it isn't useful.

@H.PWiz And it is…?

It would require that you can define arithmetic for functions from "functions from numbers to numbers" to anything. Which, you can't (meaningfully)

It's type is (Num a, Num ((a -> a) -> c)) => a -> c. And it means apply the function (+), then the function 2.

. always means the same thing

@Adám I more meant asking as wondering if those people that are "just concerned" gave examples

(unless it is used in the float literal ;-)

@dzaima Oh. No. The example usually taken up is the claim that dyadic ⍋ was hastily chosen.

I.e. it would have been better if A⍋B was A⊇⍨∘⍋B

@Adám Oh great, now I cannot unsee a face in ∘⍳∘

Also, is it possible to use @¨ in some way? I wanted to do something like 1@¨2 3 ⍳5 to return two vectors, 1 1 3 4 5 and 1 2 1 4 5

@J.Sallé that'd need hyperoperators, so no

These are sad days indeed.

@J.Sallé Wrap it in a dfn: 2 3{1@⍺⊢⍵}¨⍳5

@Adám ⍳5 -> ⊂⍳5

@ngn Yes, or ¨⍳ → ∘⍳¨

@Adám ÷/∘!¯1+⊣,+,⊢

⋄ B←÷/∘!¯1+⊣,+,⊢ ⋄ ⎕←2 B 4

@ngn
0.05

@ngn Mine is 1 byte shorter.

@Adám can we take both arguments as a single vector?

@ngn Uh, OK, but mine doesn't.

@ngn You'll like mine, btw.

@Adám here's 10 bytes but it requires both arguments in a single vector: ÷⊢/×+\!.-≡

based on the formula from the dyalog docs

@ngn Right, so that's similar to mine: ÷⊣ׯ1!.+⊢,+

@Adám ÷(⊣!+)÷+÷×

(+÷×)÷⊣!+

+÷××⊣!+

haha

The heck

i expanded the formula for the binomial in terms of factorials and instead of subtracting 1 from the factorial's arguments in the numerator and denominator, i divided out factors, and everything cancelled out nicely :)

@ngn May I add that to APLcart?

@Adám sure

so, i am going to be giving a (10 minute) introductory talk on apl at enthusiasticon, i'm a bit rusty but i am curious if people think there are things i should particularly talk about. was going to talk a bit about the history, run through an example (probably a prime number algorithm), and then talk a bit about the existing community, such as this chat room, and the other descendant languages

another (more actionable) question i had, i noticed tryapl.org has a nice display of the decomposition of tacit functions and i was wondering how to enable that in my local dyalog

@feeb When is the talk?

@feeb ]box on -trains=tree, see ]?box

@Adám some time on may 25th - i don't know the specific schedule: enthusiasticon.de/talks

@Adám i'll update you when i know when in the day it is :) it'll be very introductory because it'll be short and targeted to people that aren't at all familiar with array languages

@ngn missing on

@feeb fyi, now there's a k room too

@feeb OK, let's talk about this next week.

@Adám thanks

@ngn ahh, thank you :)

@Adám for some reason it works even without "on"

@ngn haven't worked with apl or k much at all lately, but was excited to see that shakti has an official k tutorial up, which afaik is the newest official k tutorial since like k2/k3?

@Adám ok :) let me know if i should change anything about how i am representing APL in and around the talk - don't want to step on any toes, just want to introduce people to something i've been excited about

@feeb kelas's?

@ngn: shakti.com/tutorial

ah

but the one you just linked looks great, i'll look at it :)

@feeb it bothers me very much that your name is spelled without proper capitalization in that page >.>

@J.Sallé ;)

@J.Sallé (i consider my name canonically lowercase, because that's the type of weirdo i am)

Oh, so it's on purpose? In that case, I take it back

hahahahah

@J.Sallé yup! it's been interesting to me to see when and where in my life it ends up being respected :)

Yeah, I'm trying very hard not to be the guy who points that kind of stuff out, but sometimes I just can't help myself

oh, wasn't referring to you - more my interactions with bureaucratic systems

I see, I see