@rak1507 sorting the words into order by your prime hash feels like there's a pattern in them, though it's not obvious. How about reordering the primes array so the higher values correspond with the letters with a higher Scrabble score, then would the sort be words ordered by relative Scrabble score?

not necessarily, because aaaaaaaaaa would be 2*10 (1024) whereas z would be 101, and they both have the same scrabble score of 10

what are the differences between SE and codidact?

@rak1507 good point, fixed 9 letters so there can't be z alone; but there will be that effect whatever different primes are chosen for the 1point letters

right, true

was just an example

@Jonah SE is written in .Net/Windows with SQL Server backend with closed, proprietary source code; Codidact is written in Ruby on Rails with source on Github. SE is a decade old, Codidact is a year old, give or take. Codidact was started by people who left SE (Monica Cellio, for example) to try and build a different style of community (less hostile?). There's post on their Meta with some discussions of things they will try to do differently: meta.codidact.com/?sort=score

@TessellatingHeckler Thank you.

@rak1507 You should post the merge two strings question here. Search for duplicates coming up empty...

That merge strings problem was oddly tough in J. 28 was the best I could get: Try it online! [,]}.~0>./@,(+/**/)/.@(=/|.). @Bubbler Do you see a better way?

With most of the other approaches I tried the boxing/unboxing kills you.

@Jonah I just got 23 bytes using boxes: Try it online!

I'm not going to look yet :). I'm in the middle of one last attempt. Will look at yours after.

Eh, I was on the right track but didn't quite get there: [,]}.~1+((e.<\.)~<\)i:1:. Thanks for your solution!

That's pretty close, at least :)

Equivalent of mine after some transformation is [,~]{.~((e.<\)~<\.)i.1:

and swapping args

I really wish there was a nicer way to transform left and right args in J with different verbs. You must chose between parens and dyadic hooks and @[ + @]. A lot of boilerplate either way.

@Jonah what does set -o vim do?

Who will do this in APL?

@Adám mystery

@Adám guess who did it

@Razetime Bubbler.

no it you

codegolf.stackexchange.com/a/224742/43319

boo deleted

It had a mistake.

Maybe I shouldn't try to rush for a trivial answer :(

@Razetime Let's you use vim in your terminal, basically. To browse history and edit previous commands. Life changing. One of these days I'll add a proper profile....

i just typed it in and I was wondering what it did

Now press escape, and eg k will go "up" and j will go down your previous history. / will search it. and $ will go to end of line, w advance by word, etc. all the normal vim stuff.

interesting

@Jonah The default behaviour of bash (and zsh and others) is to use Emacs keys. Is there a reason you prefer the vi keys?

Is it just because you use vi for everything else, so they keys are more natural, or is there some actual ergonomic benefits? The default emacs keys require less keypresses to do the same thing.

I just discovered that I can turn off cursor-blink in RIDE. How did I miss that for a year+?

@rak1507 Turns out it doesn't work 😂 (yet)

lol I finished that question with relatively "normie" programming

but if it works it works so ¯_(ツ)_/¯

I just thought of an extension to ⍴. I can specify one of the values in the left argument as null, and it would automatically compute the correct value based on the size of the right argument. In other words, to reformat A into a two-column array, I'd do: ⍬ 2 ⍴ A

dzaima/APL :)

This is to avoid having to type (2÷⍨≢A) 2 ⍴ A

Now I ask, is my idea naive?

@EliasMårtenson Dyalog may or may not get something similar, but instead of ⍬, it'd probably be ¯1 or ¯2 or something

also Extended

That's good too. If Dyalog is going to implement it, I want to use the same syntax.

OK, that's precisely what I wanted to do.

So does it work with any negative value, or just -1?

@EliasMårtenson some numbers may require the size to perfectly match, some might truncate or pad. Which is which is a good question

ref

I was going to have it round up.

that's ¯3 there

Hmm. I really don't like to have magic numbers for different operations (same reason I don't implement the circle operations, instead using actual words). Perhaps I should use a symbol. Like 2 3 :fill 5 ⍴ A and 2 3 :truncate 5 ⍴ A

also BQN

Ah you can do that because of the cleverness of ‿

(for APL, i of course prefer ⍬ and erroring when not exactly divisible. Shorter and I've never needed anything else)

@dzaima thanks.

I might go for that.

@Jonah post it where, SE?

@RGS :) you inspired me to try it, and I thought I had something good-ish, but mine doesn't work either, I wonder if we made the same mistake

Announcement: 2020 Problem Solving Competition – Phase II highlights has been published. Reminder: Only three months left to enter the 2021 round.

that 51 rotate is really neat

I wonder who's solution the Problem 8 one was

@rak1507 maybe :p

@Adám previous year's part 2 looks a lot more gruelling than the current one

@Razetime Yeah, we are trying to make it easier to allow more to participate.

I see.

@dzaima this is beautiful

@Adám :D cool -- thanks for letting me do it.

@ngn Is there some kind of interface with ngn/k and DyalogAPL?

@kimmolinna You mean integration of the two? No, none.

@kimmolinna no

Well, I guess you could run ngn/apl inside the HTMLRenderer object of Dyalog APL.

@Adám /k, not /apl

why you would want to interface two different array languages is beyond me

legacy?

@ngn Oh, oops.

@Adám []na might work

is there any chance of dyalog apl ever being released for freebsd? (unrelated to the above)

@rak1507 I know somehow DyalogAPL and k is mainly a database for me. I have done ipc interface for kdb+ but I'd like to use open source systems even at home.