The APL Orchard

Adám

10:57

@xpqz Fixed.

xpqz

Excellent.

Thanks.

Adám

in The Nineteenth Byte, 27 secs ago, by Adám

CMC: Given a Boolean list and a count, return the index (0 or 1-based) of the first occurrence of a run of the given count of falses. If no run of that length is found, return a recognisable value e.g. -1 or the list length or for 1-based indexing: 0 or 1+length, or null etc.

(falses=zeros)

chrispsn

11:23

@Adám {(|{$[y;1+;0*]@x}\|~x)?y} ? gives 'zero' for empty list...

Adám

@chrispsn What if list begins with enough 0s?

chrispsn

would '3' count in 0 1 0 0 0 0; 3?

Adám

Depends on the result of 0 1 0 0 0; 3

chrispsn

(ie can a 'run' ignore prior values)

 {(|{$[y;1+;0*]@x}\|~x)?y}[0 1 0 0 0;3]
2

rak1507

{⊃¯1,⍨⍸⍵⍷⍨⍺⍴0}

Adám

11:26

@rak1507 Why the ¯1,⍨?

rak1507

so that it returns ¯1 if it's not found

Adám

@chrispsn If so, then it must give 2 also when more (irrelevant) elements are at the end.

@rak1507 And if you leave it out?

rak1507

0

Adám

… which is fine!

rak1507

is it? then you can't tell between 'the run starts at index 0' and 'the run isn't found'

ah I can use ⎕IO←1

Adám

11:27

⎕IO←1 FTW!

chrispsn

 {(|{$[y;1+;~]@x}\|~x)?y}[0 1 0 0 0 1 0 1 0 0 0;3]
2

rak1507

lol

dzaima

@Adám damn horrible prototypes promoting ⎕IO←1

rak1507

{⊃⍸⍵⍷⍨⍺⍴0} then

Adám

Yup, that's what I had too.

rak1507

11:28

cool

Adám

Huh, I can save 2 (1 is easy, 2 requires either Extended or a bit of a hack) by going tacit.

rak1507

⊃∘⍸⊢⍷⍨⊣⍴0

no

Adám

@rak1507 Nope, that's not a function.

rak1507

yeah

oops!

that would be the same length...

Adám

You mean it'd be the same length if you add ⍨? Yeah, but you just need to replace that 0 to fix it!

rak1507

11:35

can't think if anything shorter than 0⍨

Adám

How about a function that returns 0?

rak1507

that's what I'm trying to think of

oh, ≡

Adám

Yup.

To save 2, you need ⎕ML←0's ⊃⍤⍸⍴∘∊⍷⊢ or Extended's ⊃⍤⍸⍴∘>⍷⊢

chrispsn

11:51

{*&~+/'y':x} appears to work in oK

Adám

I don't even…

chrispsn

y':x is acting as 'window'

Adám

Ah, so it finds the first length-N window with the sum=0.

chrispsn

yep

(empty list if no matches, which may not count as a 'value')

Adám

Clever. That's ⊃⍤⍸0=∨/ or ⊃⍤⍸⍤~∨/ in APL.

@chrispsn Why? Does *0#0 really give an empty list?

chrispsn

11:58

@Adám in oK it does - not in k9 (0), k7 (Ø), etc

k7+ has list prototypes

Adám

@chrispsn Ah, then your solution is certainly acceptable in K7.

brgal

@Adám, I am looking for this Vector article by Dan Baronet "Introduction to tree-searching in APL: Using the TicTacToe game in 3D, Dan Baronet, 17:1, p.55". I wrote to the Vector archivist email one week ago and have not yet received a reply. Do you know any one who could provide this article ? Thanks.

Adám

@brgal I think we have it at the office. I'll ask someone to scan it.

brgal

@Adám Many thanks !!!

xpqz

There are so many ks to choose from, it seems.

ngn

12:14

all you need is ngn/k :P download now for the amazing price of €0,00!

Adám

@ngn Do you take £ though?

Also, what does ngn/k do for *0#0?

ngn

@Adám post-brexit, i'm not sure :)

@Adám tio

Adám

OK, so like K7.

ngn

like all ks, i think, except k9 - the black sheep of the family

rak1507

is there a k1?

ngn

12:19

@ngn oh, and oK - it does something strange with empties

@rak1507 idk

or even k0? :) this makes me wonder, was there ever an X10? or a 6501 cpu?

Adám

@ngn X1: June 1984, X10: November 1985.

ngn

@Adám lol :)

there was a henry VII too, now i learn :)

Adám

@ngn In researching K0 I found out what "K" stands for. Anyone knows?

It appears K"0" was just called "K".

ngn

@Adám i suspect arthur makes up an answer for this every time he's asked :)

@Adám what's the explanation you found? "keys to the kingdom"? a rotated letter "A"? the easiest key to press?

Adám

12:35

@ngn "Keys to the Kingdom" ― he wasn't asked, but said it of own accord.

AndyS

16:06

@brgal Hi @brgal .. your timing was most fortuitous .. I was just about to pop into the office to swap the backup disk when Adam's request came through. If you send me ([email protected]) an email, I'll reply with the article attached as a pdf ..

Adám

@AndyS I have the email address.

rak1507

19:10

In here github.com/codereport/jsource/blob/main/LIVESTREAM_LESSONS.md it says 'John Scholes started co-dfns', is this true?

Adám

@rak1507 No, he started dfns.

rak1507

That's what I thought

Adám

But of course, without dfns, there'd be no co-dfns.

rak1507

Must be a mistake (@code_report)

Adám

Did Iverson start K?

Marshall

19:49

@Adám Maybe not, but he and Whitney spent a while at IPSA working on a prototype APL interpreter and I think that had a strong influence on A.

Adám

Oh, @Marshall I wanted to ask if you think BQN is stable enough for

17

Q: Nominations for Language of the Month, 2020 edition

We've decided to revive the Language of the Month event. We've also decided to start a new post for nominations. (You can see the old nominations post here.) So let's get to nominating! Procedure A language can be nominated for Language of the Month by posting an answer to this thread. We all vot...

Marshall

@Adám Core BQN is probably more stable than J at this point.

Adám

;-)

Adám

20:15

@rak1507 Maybe not:

> bob therriault: John did the first work on it and I believe Aaron took it to the next level

rak1507

Huh

Martin Janiczek

21:37

What a weird paper :) sigapl.org/Archives/… ('72)

That didn't age very well

Adám

lol

ngn

haha

ngn

21:55

to be fair, an epidemic is the spread in a particular region. maybe harvard in apl's case. but it didn't go pandemic.

Martin Janiczek

22:09

I'm guessing there is no recording / transcript / papers for sigapl.org/Archives/Conferences/APL2007/index.html ?

Adám

> No proceedings were published for the 2007 conference.

Hm, they planned to publish with the December issue of APL Quote Quad, but that was also the last issue ever, so maybe by then they didn't have the means…

ngn

as usual, guy steele nailed it: "the APL community [..] has been rather insular; it's been hard to make new ideas flow either out or in. The key to the survival of APL and its ideas may be increased interaction and involvement with the rest of the computer science and computing community." (from the abstract of his keynote)

which i interpret as: fix your broken terminology, make ⎕io always 0, abandon dynamic scoping, use plain text and standard editors, ..

Adám

I don't agree. K and J had already made those changes, but didn't sport increased interaction and involvement with the rest of the computer science and computing community.

rak1507

not that I disagree, but everyone will see his statement as saying 'he agrees with my ideas'

RGS

22:27

@Adám, are you available? I managed to convince @diogotito to show up here.

Adám

@diogotito Welcome to the Orchard then. How may I be of service?

RGS

@RGS turns out it was a bit more than a week.

ngn

@Adám k4 is used a lot, actually. not sure about j, and it still has weird terminology.

RGS

I told @diogotito how amazing APL is, and I also told him how you usually introduce people to APL so well :P

diogotito

Thank you!

Adám

22:30

@diogotito So, you want a quick tour of APL, or even a brief intro course?

diogotito

Well, it's been on my mind to explore an "array-oriented" language for some time and APL seemed interesting, so yes!

Adám

Aha, OK, so let's start with the array orientation aspect.

An example of that is automatic mapping of mathematical operations: ⋄ 1+(2,3,4,5) (the strange ⋄ symbol is to trigger the bot…)

Dyalog APL

@Adám
┌→──────┐
│3 4 5 6│
└~──────┘

Martin Janiczek

/me brings popcorn

RGS

@MartinJaniczek /me grabs some

Adám

22:33

Notice that we don't actually need parentheses or commas for our lists.

ngn

@MartinJaniczek haven't you watched this show multiple times already? :)

Adám

And we can map over both arguments at once: ⋄ 10 20 30 + 2 3 4 ― no explicit loop here!

Dyalog APL

@Adám
┌→───────┐
│12 23 34│
└~───────┘

Martin Janiczek

@ngn I've read a few pieces of the APL Cultivation but never saw this in flesh :)

diogotito

That reminds me of my stolen graphical calculator :D

Adám

22:35

Yeah, TI-Basic does this too.

ngn

wake me up for the moment when adam goes "tmn", it's my favourite :)

rak1507

lol

Adám

@ngn Might take a while. I'm a TMN Turtle.

Included in mathematical operations is comparison: ⋄ 3 1 4 1<1 2 3 4

Dyalog APL

@Adám
┌→──────┐
│0 1 0 1│
└~──────┘

Adám

But notice that false and true are simply the numbers 0 and 1. We'll use that soon.

We can insert a mathematical symbol between the elements of a list using / (APL uses ÷ for division): ⋄ +/1 2 3 4

Dyalog APL

22:38

@Adám 10

Adám

@diogotito With me so far?

diogotito

Aaah, so here / works like reduce or fold in some languages?

Yes!

Adám

Yes, exactly like that. It is in fact usually called "reduce" by APLers.

pitr

FYI it does tail-optimized fold

Adám

Now we can count the number of truths with ⋄ +/3 1 4 1<1 2 3 4

Dyalog APL

22:41

@Adám 2

RGS

@pitr wdym?

Adám

Just like the English language, "words" in APL apply to everything that come after them, so the comparison is done first, and then the summation.

Dyalog APL

<moon-child> @RGS foldright on a linked list (linked lists are common in functional languages) uses a bunch of stack space. foldl doesn't, and fold on arrays doesn't since they have random access. I guess that's what they're referring to

pitr

@RGS en.wikipedia.org/wiki/Tail_call

Adám

@diogotito Let's teach you another APL function: ⍳ ― the index generator. It is spelled with a Greek Iota which you can copy from here for now.

⋄ ⍳5

Dyalog APL

22:45

@Adám
┌→────────┐
│1 2 3 4 5│
└~────────┘

diogotito

ah that seems very handy already!

⋄ +/⍳100

Dyalog APL

@diogotito 5050

diogotito

Yess!

Adám

Heh, I was just about to ask "How would you compute the sum of the first 100 positive integers?"

diogotito

Oops, sorry!

RGS

22:46

@diogotito Btw, for this chat session you might want to use the JS bookmark available here abrudz.github.io/lb

Adám

@diogotito No need to apologise, you're a natural!

How would you get 11 12 13 14 15 16 17 18?

diogotito

I think I could ask Iota to give me the numbers from 1 to 8 and add 10 to that list

Adám

Right, do it!

diogotito

⋄ ⍳8 + 10

Dyalog APL

@diogotito
┌→───────────────────────────────────────────┐
│1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18│
└~───────────────────────────────────────────┘

diogotito

22:50

I'm still unsure about whether I should put spaces around operators or not

Adám

No need.

diogotito

oh

wait xD

Adám

I was just very subtle about it here:

8 mins ago, by Adám

Just like the English language, "words" in APL apply to everything that come after them, so the comparison is done first, and then the summation.

Note that there's no regular precedence order like other programming languages tend to have. Every symbol takes as right argument everything on its right.

So ⍳ sees 8 + 10 as its right (and only) argument.

diogotito

Oh, now I see!

I have to generate the list from 1 to 8 first, so maybe...

⋄ 10+⍳8

Dyalog APL

@diogotito
┌→──────────────────────┐
│11 12 13 14 15 16 17 18│
└~──────────────────────┘

Adám

22:53

Bravo. Btw, parentheses work to dictate order of execution like you're used to.

diogotito

Ok, good to know

⋄ +10⍳8

Dyalog APL

@diogotito RANK ERROR

Adám

Did you install the language bar RGS recommended?

diogotito

Just partially. I'm still copy pasting symbols right now

Adám

OK.

diogotito

22:55

Ah, I installed the wrong language. I got it now

Adám

Note that APL functions either take a single argument on their right, or in addition one on the left.

So ⋄ -10 is negation and ⋄ 10-4 is subtraction.

Dyalog APL

@Adám
¯10
6

Adám

Btw, negative numbers use a high minus, just like on that graphical calculator you used.

diogotito

⋄ -10

Dyalog APL

@diogotito ¯10

diogotito

22:56

⋄¯10

Dyalog APL

@diogotito ¯10

diogotito

⋄¯¯10

Adám

You'll see that the language bar has × too ― we use * for something else (exponentiation).

Dyalog APL

@diogotito SYNTAX ERROR

Adám

@diogotito Nice experiment. It is truly part of the number, like . is part of a non-whole number, so that won't work.

Can you explain this? ⋄ -10=-10

Dyalog APL

22:58

@Adám 0

Adám

Compare with ⋄ ¯10=¯10

Dyalog APL

@Adám 1

diogotito

Oh! Because now you used - the "word" not the sign!

10 is different from -10, so 0

-0 is 0

Adám

Yeah. We call them (primitive, i.e. built-in) functions.

And yes, that is exactly right. You're a pro already.

Transcript for