Trying to use an error guard here:

      {0::∇1⌽⍵⋄⍎⍵}'{}'
LIMIT ERROR: Expressions must finish within 10 seconds
      {0::∇ 1⌽⍵ ⋄ ⍎⍵}'{}'
       ^

in {0::∇1⌽⍵⋄⍵⊣⍎⍵} I thought if ⍵⊣⍎⍵ errored it would recurse, and otherwise it'd just return ⍵. what am I doing wrong?

{} is a function, which can't be returned from a dfn

⍵⊣⍎⍵ is still a function in this case

@Razetime You can't even catch different errors because ⍎ing both {{}} and }}{{ give syntax error

@Bubbler aha

@Razetime {0::∇1⌽⍵⋄f←⍎⍵⋄⍵} can work. I think.

I was thinking {0::∇1⌽⍵⋄⍵⊣⎕JSON⍵} but that's shorter

@Razetime I think you should post one of these even though shorter APL solutions already have been posted, as you came up with the idea, and it is a neat way to hand off the problem.

@Adám it's like the raku one yeah

@Razetime 1∘⌽⍣{∧/0≤+\⍺} is another interesting solution.

interesting

I want to partition a 2^n by 2^n matrix into quadrants.

is this possible with ⊂ or should I juse use ⌺

@Razetime It is possible with ⊂ and very hard with ⌺. Have a look at APLcart.

very hard with ⌺? thats a first

⊂ with axis should do the job I think

⌺ has a problem of boundary offsets

      {↑p∘⊂¨⍵⊂[1]⍨p←n⍴1↑⍨2÷⍨n←≢⍵}8 8⍴⍳64
┌───────────┬───────────┐
│ 1  2  3  4│ 5  6  7  8│
│ 9 10 11 12│13 14 15 16│
│17 18 19 20│21 22 23 24│
│25 26 27 28│29 30 31 32│
├───────────┼───────────┤
│33 34 35 36│37 38 39 40│
│41 42 43 44│45 46 47 48│
│49 50 51 52│53 54 55 56│
│57 58 59 60│61 62 63 64│
└───────────┴───────────┘

There we go.

Or you could try reshaping and transposing into a 4D array

@Adám much golf wow

@Bubbler Yes:

      ⊂⍤2⊢1 3 2 4⍉2 4 2 4⍴8 8⍴⍳64
┌───────────┬───────────┐
│ 1  2  3  4│ 5  6  7  8│
│ 9 10 11 12│13 14 15 16│
│17 18 19 20│21 22 23 24│
│25 26 27 28│29 30 31 32│
├───────────┼───────────┤
│33 34 35 36│37 38 39 40│
│41 42 43 44│45 46 47 48│
│49 50 51 52│53 54 55 56│
│57 58 59 60│61 62 63 64│
└───────────┴───────────┘

wow whut

I remember seeing something like this during AOC?

Behold the power of the Rank Operator combined with Dyadic Transpose!

Those two are truly the keys to the array kingdom.

dyadic transpose is my new kryptonite after figuring out imaginary numbers

The function solution would be {⊂⍤2⊢1 3 2 4⍉⍵⍴⍨2 h 2,h←2÷⍨≢⍵}

cool

2 h 2,h←2÷⍨≢⍵ can be golfed, but loses clarity.

Oh, {2⍵2⍵}2÷⍨≢⍵ is kind of nice.

⋄ {n←2÷⍨≢⍵ ⋄ 4 n n⍴(⊂⍋,n/n⌿2 2⍴⍳4)⌷,⍵}8 8⍴⍳64

multiply←{
        1 1≡⍴⍵:⍵×⍺
        a b c d ← quads ⍺
        e f g h ← quads ⍵
        p1 ← a∇f-h
        p2 ← (a+b)∇h
        p3 ← (c+d)∇e
        p4 ← d∇g-e
        p5 ← (a+d)∇e+h
        p6 ← (b-d)∇g+h
        p7 ← (a-c)∇(e+f)
        c11 ← p5+p4-p2+p6
        c12 ← p1+p2
        c21 ← p3+p4
        c22 ← p1+p5-p3-p7

        (c11,c12)⍪c21,c22
    }

quads←{⊃,/p∘⊂¨⍵⊂[1]⍨p←n⍴1↑⍨2÷⍨n←≢⍵}

:57439651           {n←2÷⍨≢⍵ ⋄ 4 n n⍴(⊂⍋,n/n⌿2 2⍴⍳4)⌷,⍵}8 8⍴⍳64
     1  2  3  4
     9 10 11 12
    17 18 19 20
    25 26 27 28

     5  6  7  8
    13 14 15 16
    21 22 23 24
    29 30 31 32

    33 34 35 36
    41 42 43 44
    49 50 51 52
    57 58 59 60

    37 38 39 40
    45 46 47 48
    53 54 55 56
    61 62 63 64

@Razetime Sure, but now you can do 2 2⍴⊂⍤2

Just wanted to demonstrate a radically different "flat" approach :)

@Adám needs , before then

@Razetime In production code, I highly recommend parenthesising multiple assignments: (a b c d)←…

I wonder if you can do (2 2⍴a b c d)←2 2⍴2

@Razetime Why would you?

@Adám was just learning this in algorithms class and thought it'd be convenient

to try in APL

@Adám funny reasons

@Razetime Ah, you actually don't want a 2×2 matrix, but a list of quads.

@Adám yeppo

@Adám I KNEW that was there somewhere but couldn't recall it - excellent!

@Razetime Then simply {⊂⍤2⊢2 1 3⍉⍵⍴⍨4,2 2÷⍨≢⍵}

@MortenKromberg I don't remember such by heart either. I just constructed it.

i can't tell what's going on there lol

strassen's algorithm pretty cool

@Adám Indeed, I meant recall the technique

@Razetime Reshape to 4 (n÷2) (n÷2) then swap first two axes and enclose the trailing matrices.

There are some real uses of dyadic transpose

@Adám umm

      {⊂⍤2⊢2 1 3⍉⍵⍴⍨4,2 2÷⍨≢⍵}4 4⍴⍳16
┌─────┬─────┐
│ 1  2│ 3  4│
│ 5  6│ 7  8│
│ 9 10│11 12│
│13 14│15 16│
└─────┴─────┘

@MortenKromberg Right, it is almost always: reshape, dyadic transpose, apply function with rank, post-process. Though sometimes, some steps can be skipped or simplified, e.g. using monadic transpose.

@Razetime Oops.

⍋

Time to introduce ⍋⍛⍉

As I understand it, the reason for the design of ⍉ being that it allows you to select diagonals by giving one or more dimensions equal rank, whereas simply selecting dimensions from the right would not allow that.

... so it is the more complete/powerful of the two options

I so want ⍛

Anyway, dyadic ⍉ is the perfect example of why you need a REPL :D

It can also be used to "fix" ⌷ as ⊂⍛⌷

@Razetime Ah right, I realise now that we do need to go via 4 axes, as those are the actual dimensions.

I am in head empty mode, so I'm gonna just bookmark this conversation to read later

⊢⍤ works around a design flaw in / et al. ⊂⍛ works around a design flaw in ⌷.

@Adám I'm not sure "fixes" is the right word, perhaps "works around" would be better, but I get your point.

@MortenKromberg Fixed :-)

@Adám @Razetime I know ≢ is the cool new thing, but 2 2÷⍨≢⍵ is 2÷⍨⍴⍵

@MortenKromberg Doh.

@MortenKromberg And that's probably more correct if two dimensions differ

@Razetime {⊂⍤2⊢1 3 2 4⍉⍵⍴⍨,2,2÷⍨⍪⍴⍵}

@MortenKromberg cool and good

@kimmolinna Welcome back. It has been a while.

@adam Thanks. There has been quite many things on board.

@Adam I have a problem with ]LINK shall I send you an email or ...

@kimmolinna Either way.

@adam I will get Error: VALUE ERROR: Undefined name: warmsg with Getting started Example

@adam when I try to do - ]LINK.Create stats /tmp/stats -source=ns

I will use it on Linux with just copied files from Github

@kimmolinna Good morning! I'll take a look. Stand by.

Thanks.

For the record, can you confirm ⎕SE.Link.Version?

2.1.0-beta61

Dyalog versino 18.0 under Linux?

Dyalog APL/S-64 Version 18.0.39712

It was strange that I got a warning about missing lib* and then I updated LINK and deleted .dyalog to get a fresh start.

Rats... it works fine under Windows, I will actually need to fire up a Linux box :(. Before you retort "typical", know that the main Link developer, uses Linux as his primary environment :).

Sorry guys. My bad. .NetCore wasn't installed and your code just missed one n.

I installed kubuntu again yesterday...

@Adám Ah, no, sorry -- my comment was meant to mean "check that out; it is documented and works (once told where to look)" :)

I know I'm annoying, but who on earth decided the timing for dyalog.com/apl-seeds-user-meetings/aplseeds21.htm, right in the middle of the day doesn't seem like a great time for most people

@kimmolinna Your bad, except for the typo in the Link code (warmsg rather than warnmsg) which meant that the warning turned into a crash :(.

@rak1507 Whose day? We do pretty much all our events at this time so that it is possible to attend from Bangalore to Seattle without TOO much grief.

... with apologies to our potential viewers on Hawaii, NZ, Japan etc

@MortenKromberg it being on a weekday means that most people with school or work probably won't be able to attend

14:00 is 06:00 on the American West Coast, and 18:00 is 3 AM in Japan and Korea. I guess we could have made it one hour earlier to include more Australians.

@rak1507 There is no perfect time, fortunately we will record it so anyone can watch it after the event. Last I heard, about 160 people had signed up to attend.

We just hit 170 today :-)

Do you have stats for the experience levels?

@rak1507 This claims Wednesday is best.

@rak1507 Yes:

wow, that's surprising, fair play, looks like you're right

Do you want per-country stats too?

sure

the more stats the merrier

India 7!!!!!!

I actually expected more from India.

US > UK, surprising considering dyalog HQ is in the UK

but I guess the big businesses are all in the US, so it makes some sense

Wait, why does it matter where HQ is?

I thought the HQ being in the UK meant most of the target audience were also in the UK

Fun fact, <2% of Dyalog's revenue is generated in the UK

wow, I assume ~lots is from the US then?

The bulk is more or less evenly split between US and EU

interesting

@rak1507 We kind of have 2 HQs, the other being in Denmark, yet we only have 2 Danish registrations.

I suppose at the moment with everything being remote anyway, it doesn't matter where they are

Well, we've kind of intentionally aimed this meeting at "new users", which may have made most of the big customers instruct their developers to watch the recordings rather than "interfere". I think I will send out a few reminders.

@Adám oh, there are more?

I don't know too many people who have heard of APL here

@Razetime Not a lot of people, but there are a few outsourcing companies in India that do APL work for companies in Europe and the US. Also, speakers from Dyalog have regularly been on the programme at functionalconf.com in Bangalore, every year that it has been held.

Waiting for next functionalconf

@Adám for the longest common subsequence question, none of the examples use alphabets other than ACTG, but this isn't guaranteed, right? (oops, one of them does)

@Bubbler You can use ⍉⍣¯1.

      ⎕IO←0
      ⍴ 1 3 2 0 4 ⍉⍣¯1 ⍳0 1 2 3 4
1 3 2 0 4

@Marshall Ooh, nice. Didn't know/think of that.

mortens description of dyadic transpose made me realise exactly why all my uses of dyadic transpose so far have been something like 'iterate through all possible left arguments and pick the one that works' lol

@rak1507 The way I do it is writing ⍳rank and then I move the indices to the positions I want.

What clicked for me (in the shower where I do all my best thinking) was visualising the 3-d case with an imaginary "camera" position, looking at the data from different directions, thinking how the coordinate system changed. Oviously harder for higher ranks than that..

@Bubbler I'd make that 5D, but yeah…

I think that was the day when AoC threw out 3D and 4D GoL

@user5392393 Hi theuberdog. If you want to participate here, email access@apl.chat

@Bubbler :D well, quite so

Is it ok to discuss old (that is, not current edition) Dyalog comp problems here?

don't see any reason why not

@xpqz For sure. Any APL usage is on-topic

Problem 6 from last year's phase 2 -- the JSON mail-merge thing... I have a filthy regex thing that does the job (gist.github.com/xpqz/2cab73665d6bc4b20629a65275b4a62d) -- but is there an array solution?

@xpqz Ah, yes. I keep meaning to publish the cream of the crop from last year. Here's the solution I selected for that collection:

Merge←{

   ⍝ Cut the template on '@', then replace every other part
   ⍝ by its value in the JSON namespace.

   templateFile←⍺
   jsonFile←⍵
   template←⊃⎕NGET templateFile
   ns←⎕JSON⊃⎕NGET jsonFile

   getValue←{
       0=⍴⍵:,'@'   ⍝ '@@'         → ,'@'
       6::'???'    ⍝ ~⍵∊ns.⎕NL ¯2 → '???'
       ⍕ns⍎⍵       ⍝  ⍵∊ns.⎕NL ¯2 → ⍕ns.⍵
   }
   ∊getValue¨@(⍴⍴1 0⍨)'@'(1↓¨=⊂⊢)template

}

Thanks -- I shall spend some time with that.

Why does the solution for "is a vector strictly increasing" involve a sort on aplcart? (⍳∘≢≡⌽∘⍒)N

Sort is probably going to be O(n log n) which is worse than the ideal O(n)

It isn't sort, only grade. Though that may have worse performance than sort…

Grading still involves a sorting operation, no?

@EliasMårtenson Questions is, should it be changed to ∧/2≤/⊢ or should I change N to Y?

      ]runtime -c '∧/2</a' '(⍳∘≢≡⌽∘⍒)a'

  ∧/2</a     → 2.0E¯5 |     0%
  (⍳∘≢≡⌽∘⍒)a → 1.9E¯3 | +9795% ⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕⎕

I'd write it as ×/0> 2 -/ N

@Adám Or that. :-)

@Adám change to ∧/2≤/⊢ imo

(can a strictly increasing array have equal elements?)

Sorry, </ for strictly.

@EliasMårtenson ⍋ is the only primitive in Dyalog that can compare whole arrays, and that solution works on nested arguments

@EliasMårtenson Oh, now I realise the correct answer: Notice that it says N and not Nv.

@Adám I think you missed that the sum of values in your code has to be a multiple of 1729 in this answer codegolf.stackexchange.com/a/221330/95516

and the ngn/k one

@rak1507 Oh, man. Yes.

@xpqz Discussing wishes for future versions of existing array languages here is also "on topic". In my opinion, the point at which it is appropriate to consider using alternative forums is when discussing proposals which depart radically from existing APL or J.

@MortenKromberg I was more concerned with spoiling the fun for others doing the problems.

@xpqz OK, that is of course always a problem!

@xpqz I think it's been sufficiently long since that anyone here will have already seen the solutions

Yes, I think so too, on balance.

It will also disappear into chat transcript oblivion fairly quickly.

Sorry, offtopic question, but which application do you use in Android to read these messages nowadays? The Stack Exchange App is not in Google Play anymore.

@Adám ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||+/⎕UCS⍞ fixed :P

@kimmolinna the app never had chat, so the web chat interface is, and was, your only option. The app did do push notifications though

@dzaima Ok. Thanks for info.

|∪|↑,⊃+/⎕UCS⍞ slightly shorter

@rak1507 I already edited to a 10-byter.

ah, comments :(

no fun

Sorry, excuse me.

@MortenKromberg BQN is an existing array language that is closer to J than APL, but Adam says that discussing its design is off topic. So I don't think this is accurate.

And the feature that was in question at the time, Nothing (·), is just a syntactic approach to J's cap ([:). It's not a radical departure from APL notation.

@Marshall Morten was just expressing his opinion. Though if you include changing established spelling in "radical", then BQN is excluded, despite being similar to J and APL.

@Adám It seemed to me that he wanted to describe the actual rules for this forum. Do you agree with me that the rules aren't as he says?

This was the only official announcement I made.

Is ELI on topic? It's also spelled differently from J or APL.

Regarding discussion of BQN's Nothing. Was it on topic or not? How are we expected to know when you're acting as moderator if that doesn't qualify as "official"?

@Marshall That was in response to a flood of joke messages about the word nothing, simulated backspaces appearing to remove characters from messages, etc. It was just noise.

You proceeded to make some comments indicating that discussing BQN's design is off topic. Can you please clarify what the actual rules are?

@Marshall It is the in-depth language design discussions I want to get rid of, because they run the risk of drowning normal questions etc. We used like 300 emails internally at Dyalog to discuss Total Array Ordering. That would have been inappropriate here too.

@Marshall No I can't.

I think the community is rather concerned that what's allowed on this forum is determined by a single Dyalog employee making judgment calls.

Smaller discussions about existing, possible, desired, potential, etc. design are fine. Flooding the room with long arguments about a single detail or a a few connected details isn't OK.

But there are no hard and fast rules. Try to be nice.

As was pointed out at the time, please say that instead of declaring the content "off topic". What you are now saying is that the topic was acceptable but the approach was not.

@Marshall Right, so I was missing one term: "in-depth".

@Adám "In-depth" is not a topic, it's an approach.

Exactly.

Actually, let's simply replace "Non-APL" with "In-depth".

K being off-topic is something I've considered for a while. I asked ngn about it long before these discussions started.

@Marshall I should also point out that I'm only human, and of limited ability too, and I don't have a lot of experience running APL chat rooms. Please forgive me for not always getting things right on my first attempt.

@Adám I'm saying that it is not correct to prescribe how discussion should be approached as part of the room topic. If necessary, the topic should clarify what can be considered APL.

@Adám "in-depth" is just something that any conversation can convert to arbitrarily without knowing it ahead of time. As switching rooms precisely at the point when it begins to classify as "in-depth" is incredibly impossible (getting everyone to move to a different chat system, somehow referencing previous messages, actually figuring out at what point to do that), the only safe thing is to never talk in the orchard

@Marshall Right, but by removing "APL" from the topic, we avoid that issue.

@Adám I mean, it hardly matters to me. The new forums have BQNbot, I'm abandoning this place as fast as I can manage.

No you're not. You're here ;-)

I think the type of content you want to be in this room is "beginner-friendly" at whatever cost

And I really appreciate your contributions here.

@dzaima Feel free to think so. It isn't the case, though.

(is this discussion about on-topic-ness getting too in-depth? should we move to somewhere else?)

Probably.

@dzaima Who says it has to be an exact science?

I agree that it is an issue that break-out rooms cannot easily be made. (Well, they can, but then many people cannot participate there.)

The various other chat places, IRC, Discord, TopAnswers, do not have this limitation.

However, this room probably has better discoverability.

@Adám TA is probably the only place in which it could be acceptable to just spam new rooms as needed (just because it isn't used for anything else). Everywhere else is just fine with various degrees of off-topic-ness everywhere. You're the exception

Hence my request to take (and possibly link) discussions elsewhere.

@dzaima "various degrees of off-topic-ness"? But some off-topic-ness is OK here too. You'll occasionally find me chatting about religion or human language.

@Adám what else can it be? noone could've known that the nothing discussion would anger you to the level it did when it began

@dzaima And about spamming rooms, yes, that's why I specifically mentioned TA.

I don't think spamming new rooms is good in general

@dzaima Devil's advocate: On SE, long discussions in comments on posts should continue in chat, but at what exact point. Nobody can know in advance. Best not to comment at all? Surely not. We're humans trying to do our best.

@Adám noone gets mad when you get it wrong though

Did I yell at anyone? Did I kick or suspend anyone?

you do whatever makes sense to you. Not to some arbitrary single moderator

Anyone joining the BAA webinar today? Someone reached out to me asking if it's still on schedule, apparantly only 3-4 people in it

@JoshD It starts in 47 minutes

@JoshD Another 45 mins. USA and EU are one hour less off than usual.

@Adám you hard-stopped the discussion at least. No SE moderator would force you to stop your ongoing discussion in comments

Ah, thank you

@dzaima No I didn't. I do have the powers to hard-stop discussion, but I didn't use them.

@Adám and you still wouldn't stop it if we went on for another 2 hours?

"You've been warned." from a moderator indicates intent to use those powers. Please don't say that if you're not acting as a mod.

@Marshall OK.

@dzaima Out of principle, I refuse to answer questions about what I would have done if things had progressed differently than they factually did. I consider it a waste of time, because no answer can be verified.

@Adám y'know, it's kind of very important to know if there's ever a point when we'd be forced out of the room. If there's not, nothing has changed, and, as long as the initial discussion starts as on-topic, there's no fear of being kicked out later. If there is, well, then this place is useless for any discussion that has any potential in getting into the "kicked for off-topic" category at any later point

my whole problem here is that it's fine to start a discussion, but maybe not to finish it. That's a horrible way to do conversations

@dzaima Obviously, it is possible to be forced out due to actual spam or other "not nice" behaviour.

@Adám well yeah. The "gray zone" is the more important part here, and we can't even know what to expect

@dzaima Don't be afraid to start any on-topic discussion. Please try to be sensitive to whether the discussion has progressed to a point where others could feel that they don't have a space here, even though they really should have one.

@Adám "Don't be afraid" I'm never gonna not be afraid if there's any possibility of being kicked out just for the conversation going "in-depth"!

"others could feel that they don't have a space here" that sounds awfully like my '"beginner-friendly" at whatever cost'

@dzaima Why not? I'm not going to kick anyone out just because they got carried away with an interesting on-topic discussion. I might tell you that I you should take it elsewhere. However, if you start spamming useless noise (and some people did, with backspace and "nothing" jokes), then you can rightfully fear being kicked out. I'll always warn first, though.

@dzaima the whole problem is: what to do after we get kicked out? just stop the discussion completely? create a new SE room (which noone will ever see)? move to a different platform? (a hassle for everyone involved)

@dzaima No, an APL expert asking for tips is 100% OK too. I don't know where you get this "beginner" thing from.

@Adám an expert asking for tips is still beginner friendly i'd say, and if that discussion would go on for a while, i'd guess you would start not liking it too

@dzaima By "kicked out" I assume you mean "asked politely to take it elsewhere". You could post Let's continue this [on {platorm}](link). I agree that it is a hassle, but unfortunately, we don't have much influence on this platform's technical side.

@dzaima Why do you guess that?

I've had discussions (and lessons, etc.) here that went on for many hours. As long as there is room for others to come in and participate with subjects that are the main goals of the room, then that's fine.

@Adám for most conversations, the hassle would be greater than the need for continuing the conversation. So, effectively, you're just shutting down conversations you don't like

if the platform has an issue, you don't just pretend it doesn't exist, but try to work with it, or move somewhere else

@dzaima It isn't about what I like. It is about keeping the room accessible to others.

@dzaima I'm happy to hear what you have to suggest. How do we keep the room welcoming and still allow in-depth language discussions?

@Adám anyone who's been in any chatroom for any period of time would be just fine with discussions they can't participate in. The others? They're the begginerss i'm talking about

SE is also one of the best platforms for multiple concurrent discussions, so I just don't see any issue

@dzaima Oh, so you mean chat newbies, not APL newbies? That wasn't clear to me. I'm sorry.

"if the platform has an issue, you don't just pretend it doesn't exist"; "I just don't see any issue". Lol.

@Adám well, I mean both really. APL newbies will often be SE chat newbies

@Adám The "issue" in the first is that there's cost in creating new rooms. The "issue" in the second is that "in-depth" discussions are somehow bad.

in depth != unnecessary noise

I agree there was unnecessary noise and jokes on the nothing discussion, but this pretty clearly said nothing about that specifically, and said specifically that design discussion is noise

yeah, there has been unnecessary stuff, but there's also been in depth discussion, and I think the second one shouldn't be banned

Right. In the heat, I looked at the two as one thing. Some of the participants used typographic jokes in their language design discussion. The flooding was what bothered me. Unfortunately, I ended up dismissing design discussion in an attempt to limit noise.

recommending creating a whole room just for nothing jokes is a bit weird though

Don't diss No Place man. "No Place is perfect for everyone!"

Can we continue this discussion at another time?

this also says "language design is off-topic", which isn't noist as far as i'd say

Announcement: BAA Webinar in a couple of minutes. Passcode: ×/1920 12 17

^ join if you want to see old men yell at clouds

@dzaima (well, "Non-APL language design" more specifically; which i'd say is much more harmful - disallowing discussing competitors is a pretty pro-APL move)

@rak1507 That could be taken as a "not nice" message.

probably

@dzaima Am I allowed to reconsider, or are we forever bound by my previous statements?

@Adám if you want to. BQN has already mostly moved away though, so you'll get you're purer APL experience anyways

OK, so language design discussions are on-topic again, but please do try to keep the place welcoming, and above all, always be nice.

∧/⊃=⊢ seems faster than ∧/2=/⊢, interesting

@user14984631 Hi Abhishek Kumar. If you want to chat here, email access@apl.chat

@Adám going back to the "Merge" competition problem we discussed before, I can't see quite how the version you posted can work correctly. If the template file doesn't begin with a template, the bit before the first template will be skipped, rather than left in situ. Unless the competition data can be assumed to always lead with a template, but that's not how I read the task.

      '@'(1↓¨=⊂⊢)'aaaa @bbb@ ccc @@ @ddd@'
┌───┬─────┬┬─┬───┬┐
│bbb│ ccc ││ │ddd││
└───┴─────┴┴─┴───┴┘

it should really be ≠⊆⊢ or something

I think with a leading non-template, the neat 1 0 1 0 1 0.... assumption falls apart, too.