@dzaima How do I split on spaces in dzaima/APL?

@Adám ' '(1↓¨=⊂,) or something

Any chance for ⊆?

@Adám i don't like it merging consecutive separators

..and its weird ascending thing that it does

I have an expression to extract from a matrix a submatrix containing all rows for which column 4 value equals a given value: {(matrix[;4]=⍵)⌿matrix}. I can pass a vector to this expression and it produces a vector of matrices, which is what I want: {(matrix[;4]=⍵)⌿matrix}¨ 1 2. Now I'd like to replace the matrix reference by a left parameter ⍺. But matrix {(⍺[;4]=⍵)⌿⍺}¨ 1 2 produces a rank error. What is wrong here?

@RudiAngela the ¨ will apply to the left argument too

so you want do ⊂ the left arg

@Adám he's 24 now, but started looking at APL a few years ago. He was a Phase 1 winner this year (dyalog.com/student-competition.htm) and also wrote a significant portion of aplwiki.com/wiki/Stencil . I can only look on with pride :-) Anyway, we both enjoyed watching various APL videos, from the classic 1975 Demonstration video, through various Dyalog videos, to the more recent Code Report ones too.

@RudiAngela Either what dzaima wrote or alternatively insert ∘ on the left of {

@qmacro Oh nice.

@dzaima Thanks. But where do I place that ⊂?

@RudiAngela (⊂matrix) {(⍺[;4]=⍵)⌿⍺}¨ 1 2

@Adám Thanks. It works, magically (i.e. I don't know what I'm doing here).

@RudiAngela You're binding the matrix to the dyadic dfn to form a single monadic derived function, which you are then applying to each of the numbers.

@Adám My previous reaction was to your jot suggestion.

@Adám Ah yes of course: need to place in parens.

@RudiAngela Yes, I could see that because you used the reply functionality (thanks!) and my explanation is indeed for that.

@Adám OIC. Makes sense.

@qmacro I wish I had known about APL when I was 24 ! The closest to APL that I knew was Mathematica. But, I believe that APL is much much better. I'd be proud of my son too if he was making steps helping a nice & important community as APL's :) I hope he keeps it up.

imho APL should be taught to every math & physics department around the globe along with some language like Fortran or C/C++ and Mathematica (for symbolic computations, such as symbolic integral evaluations).

It'd be nice if in the future a symbolic math library like sympy would be created for APL.

@user82382 Hi Guys Math! If you want to participate here, simply email access@apl.chat

@klao Hi there. Interested in APL?

@Adám I saw that there's a Springer book out there about creating an APL compiler. I think Haskell has an interpreter and a compiler. Does Dyalog have any plans into creating a compiler for its APL too ? :)

@milia Well, we have an experimental built-in compiler and are sponsoring Co-dfns.

@Adám thanks, I'll surely check it out :)

@Adám this is gold material: bonfire.com/store/arcfide :D

christmas is soon.... wonder if santa could get me something :P

@milia I can't testify about all the products there, but I can assure you that the first three items are greater than average.

@Adám as I just mentioned right now to Aaron, I'll surely honor his shop, as I find the idea of going around with some APL code on my chest a great conversation starter :)

and hey, if it's of top quality, as it looks like it, it's a win-win

@milia I hope you got my joke.

Dyalog should sell merch with the hexagon icon, and the ⊃∨∧⌊○⊂ logo etc.

oh right

the joke must be related to the code of the first three...

I got this at least :O

:P

@Adám yep, having a Dyalog sweater with some nice APL code on it would be nice too

Maybe a shirt that says ⊣⌿ on the front and ⊢⌿ on the back. That'd be cool.

@Adám Wait a minute, how is Co-dfns only 17 lines long?

I think it is only the core code.

easy, it isn't :P ^

Ah

damn, I think Aaron's sweaters are like a test of being accepted in the community. Like in special forces. You have to go through hell -they say- to get your green beret. I guess I'll go study more APL so that I understand the fnc on the sweater. Or else I wouldn't call me worthy of wearing it xd

thankfully xpqz.github.io/learnapl/intro.html is online again :)

Clothing claiming that 42=6×3+4 – how's that for a conversation starter?

@Adám hahahaha

Or that 1 = 6÷2×(1+2), given that memes about its ambiguity have been going around Reddit

I wonder what percentage of people with co-dfns tshirts can claim to fully understand it, probably not many

@rues no, it is 6÷2(1+2) that has the ambiguity

which is obviously 3 2

no

Only using execute.

Gotcha, thanks

yuck

Was just curious :)

No, wait, that's not necessarily right.

You might well be able to.

:O!

hm?

You'll need execute if you need to access global names at runtime, but not otherwise.

@rak1507 Oh yeah, forgot that was valid APL

Stranding is weird

Heh, 3 2 ≡ 6÷2(1+2) would be nice.

too weird to be interesting to normal people but too obvious to be interesting to aplers

Or something like

I'm surprised 2(1+2) doesn't do multiplication, considering APL was originally used as a blackboard/paper math notation. I think the multiplication being implicit is why most people get confused with 6÷2(1+2)

¯1 ≡ det (2 2 ⍴ 0 1 1 0)

@FawnLocke No no, Iverson wanted to make an unambiguous and explicit notation.

:)

@milia But det isn't really APL.

@rues would work better with the galaxy brain format

@Adám Was the stranding designed by Iverson himself or added later?

@rues Goes into the Memes section of APL Wiki?

@rak1507 True, true

@Adám Yeah, of course. Just an obvious example where iverson didn't follow mathematic tradition.

@rues Iverson was opposed to stranding!

@Adám There's a memes section? Nice

I think Iverson wanted (1, 2, 3) notation originally

1=-*○⊕1 is nice, but not yet implemented :-(

@FawnLocke Not just wanted; had.

@Adám Oh, was it John Scholes then? I think someone linked to something about him implementing it

Which implementation?

Or, was it before an impl

@rues No, John Scholes was much after that. I don't remember who it was, but it was only of the first implementers. It came up in a recent APL Campfire.

@rues apl.wiki/Humour#Memes

@FawnLocke The very first. Didn't have a name yet, I think, or maybe it wwas PAT or IVSYS/7090.

APL Wiki seems to be more dedicated to humor than information sometimes :P

Thanks

@rues Humour is important, and the early APL giants had plenty of it.

Is Alpha and Omega on the humour page?

@FawnLocke No, but feel free to add them if you think it fitting. Though, I don't think it was ever humorous meant. Originally, the symbols, together with their overlined siblings, were dyadic functions.

it's unintentional humour

Why is it even funny?

Out of context it can look like APL is a religion :)

@dzaima I wonder why Judaism isn't listed first there…

@Adám well, what if det is simply a short name for some complicated APL dfn? :)

@FawnLocke some have said that it is. I think Dijkstra in one of his emails.

@milia That'd work in any language.

@Adám You edited that page, didn’t you?

@Adám APL is the only language that has a rho symbol though afaik :)

@milia jsoftware.com/papers/Dijkstra_Letter.htm I think this is the source

@FawnLocke exactly.

Quite a funny interaction :)

Jsoftware has tons of interesting material to read.

@milia Sure, but why not -1 == det([[0,1],[1,0]]) and say it is JavaScript or Python or somesuch?

I found also the book of Roger on the interpreter of J in C pretty interesting find. I have yet to finish reading it, and then digesting it.

@Adám because my proposal is classy, whereas yours -sorry to say- makes my eyes bleed...

@FawnLocke Hm, that should be added to apl.wiki/Edsger_W._Dijkstra

@milia For a two-row matrix, it is simply -/×⌿0 1⌽m

> Dijkstra didn't hate APL specifically: he hated every language other than ALGOL 60

@Adám It's in the external links section, saved me an edit :)

@FawnLocke Yup, I just notices, but why isn't it more front and centre?

I was also thinking that

:59728046 They are about the same, but hilarious that the original uses ⎕IO←0 for footnotes, while Vector's uses ⎕IO←1

@FawnLocke The Wiki is a free-for-all, go! Edit away!

Lol

@Adám taking notes !

@rues lol!

@rues What?

@FawnLocke Even stated in the latest Array Cast.

@Adám He couldn't have hated every single language other than ALGOL, could he?

Why not?

It even mentions him saying something positive about LISP later

@Adám Because there were so many languages around, even in the 60s

Surely a few others interested him, or at least weren't bad enough by his standards that he didn't hate them

¯\_(⍨)_/¯

I wander what Dijkstra thought of Fortran.

Probably a dozen articles on it :)

@FawnLocke I'll google then, thx :)

Looks like he hated that too

> FORTRAN --"the infantile disorder"--, by now nearly 20 years old, is hopelessly inadequate for whatever computer application you have in mind today: it is now too clumsy, too risky, and too expensive to use.

@rues I begin to like the man then.

@FawnLocke still true at the most of it :p

*for

I can't speak to the virtues of Fortran, I have heard that modern Fortran is quite nice to work with, compared to it's older versions.

@FawnLocke more pleasant yes, maybe...but...still fortran

:)

F77 was horrible...I was introduced in it around 2001 ...

Maybe when we settle on an APL logo, I'll open a merch store.

@Adám how's that going?

If you can share ofc

“The sooner we can forget that FORTRAN ever existed, the better, for as a vehicle of thought it is not longer adequate: it wastes our brainpower, and it is too risky and therefore too expensive to use.” (862).

Who's gonna tell them the bad news