@Adám Is there an obvious way other than binary search to get the largest and smallest integers, before we overflow into floats?

Oh, it's just the range -2*31 to ¯1+2*31.

       e1←'hi' ⋄ 1e1e1.1
 10  hi  0.1

Number parsing for fun an profit?

@B.Wilson You're clearly not having enough fun:

      J←+
      1.J.J.J.J.1
1.1
      1.J.J.J.1
1.1
      1.J.J.1
1.1
      1.J.1
1J0.1
      1.1
1.1

@Adám LGTM. Merged.

APLX is real only right?

(the wiki says it has floats but then lists + as conjugate)

Yeah, that's a mistake.

Ooh, this is interesting. APL2 has + as strict conjugate, APLX as strict numeric identity. Dyalog's + isn't just conjugate but is actually "negate imaginary part", because non-numbers, are in the domain, bug having no imaginary part, they are returned as-is.

does APL2 not have ⊖?

Sure it does. Just tried it.

interesting, can't find it in the docs you sent me

what were the primitives that needed to be split? I currently remember compress/replicate but I think there were others

@RubenVerg apl2lrm page 228

oh, I was looking at LS

@RubenVerg oh circular also needed to be split right?

@RubenVerg You can look at the transcript.

oh ¯ is a valid identifier charactter in APL2, interesting

Yes, but not as the first char — and this applies to _ too, whereas Dyalog allows leading _s.

does APL2 support ⎕⍞ for IO? quickly browsing through the docs I couldn't find them and Ctrl F doesn't work on the language reference

alternatively, can I download APL2 from somwhere and try it out myself?

hmm, looks like ⎕← is supported, can't find anything about the other three options though

(⎕, ⍞, ⍞←)

@Marshall Didn't seem to be anything definite, no announcements that I could see. Vector 5-1 (july 88) has news of beta program and 5-2 (nov 88) performance success, and what seems to be demos at APL-ication conference. 5-3 (jan 89) has a listing in product guide section as you'd already found

@RubenVerg Do you look at the ToC?

@Adám apparently not :) oops

I think I have everything for APL2 then - I'll push the new data soon

This will be interesting!

I mean, I'll surely have missed a bunch of things :) reading documentation is hard

I also wasn't sure how to handle Replicate/Expand being operators, currently I just duplicated the entries and changed the description to "operand" instead of "left argument"

Sounds right.

wait, are APL2's format by specification and Dyalog's dyadic format equivalent?

i currently have them as separate entries but it might've been a mistake

looks like it's the same

Yes, I think that's standard.

alright, I deployed the APL2 glyphs

oh wait I forgot about the ^/∧ thing

should I mark the dialects that support ^, or just APL2?

up to you, but the others (except dzaima) do support it.

I mean, I guess I mark j and J both for dialects with both, so I'll mark ^ and ∧ both as well

oh but then I have to mark all the extensions on ^ as well

yeah I'll just mark it for APL2, secondary glyphs aren't really a thing I've been considering

ok, pushed the fix

@RubenVerg Beastly work! Where's this repo, btw?

@Adám is this the only difference? (GNU APL is next on the list of things to add)

@B.Wilson RubenVerg/omnibar

btw I'm missing a favicon, any ideas on what I should use?

@RubenVerg Maybe ⍰

@RubenVerg I believe so.

@Adám oh, great idea!

Do you want to distinguish between monadic + being a numbers-only function and accepting all data?

@RubenVerg Did you know you can actually use a character as favicon, by embedding it in an svg?

would it render properly on a system without an APL font? or can I embed the font as well?

@Adám probably not, seems too granular

there's other stuff I group together that technically isn't the same — dzaima's ., for example

@RubenVerg I think it'll just pick up the font if you specify it. You're hosting BQN386, right? (Btw, why not APL386, as that'll give you NARS2000's derivative glyph?)

@RubenVerg And monadic ≡ gives different results on un-evenly nested data with Dyalog vs APL2.

@Adám I'm currently not hosting it, just pulling it from dzaima's website

@RubenVerg Hard to distinguish the two. Maybe add border:1px solid silver;border-radius:.1em which also helps distinguishing . from ∙

@RubenVerg Uh, and you're also using cdn.jsdelivr.net — I believe both are actually illegal many places.

illegal?

also I'm not intentionally using jsdelivr, must be something from svelte (or the bootstrap library maybe?)

Yes, because when people visit your page, their IP address will be forwarded to a 3rd party. This means that the 3rd party can register that this particular person (IP address is considered PII) visited your particular site. You can't give away that kind of info without explicit consent.

Maybe also padding:.1ex to avoid underscored glyphs touching the border.

@RubenVerg Missing E

Heh, " will need duplication when you add APLX, as APLX's " is like '.

@Adám what? so CDNs are illegal in many places? never heard of that

@Adám thanks, I'll add it

@RubenVerg Your description of ?y is very explicit. If you keep it that way, you need to split ?0 off, since APL2 doesn't support it. Also, it seems there's some kind of typo: "integeme([R], s)"

@Adám find and replace went wrong! does it happen anywhere else?

@Adám hmm, suggestions for a less explicit description?

@RubenVerg At least unless you have an explicit privacy and data policy on your site, and that the contract you sign with the CDN has the appropriate stipulations:

The safe thing is to self-host everything.

@RubenVerg "Random numbers as specified by the argument" which also fits Vision as it allows way more choices in the specification.

@Adám a lot of documentation suggests using cdns, I wonder how many websites are non-gdpr-conformant

I'll look into what's being pulled from a CDN and try to host it myself

Most websites aren't GDPR-compliant, and that includes most websites for companies that offer paid assistance with becoming GDPR compliant!

oh, svelte really doesn't like me locally importing Bootstrap

should be fixed now, hope I didn't break anything

I love svelte but sometimes it's so annoying

I'm tempted to switch to plain JS but I don't want to do this all over again

@Silas Thanks for looking! Seems reasonable enough to put the full release at 1989. Apparently there are proceedings for APL-ication, wonder how hard those are to find...

no luck there either - think must've been in one of the product slots

As in, you don't have the proceedings or the proceedings don't have it?

the proceedings don't have it

APL-ication should have an entry on BAA's page, like BAPLA.

Do you have the place and dates?

@Adám hmm, maybe this one should be split

not sure

@Adám great suggestion, I'll add this

deployed it, not fully convinced about how it looks though

@RubenVerg Way too much padding, and outline is too dark/prominent.

Also seems like the char is offset below the centre of the border.

@Marshall yep - proceedings say University of Kent, 28th-30th September 1988

Thanks, was just about to add it without that info!

Here it is.

@RubenVerg Compare my vision with what you have now.

hmm, that definitely looks better

is it the same code you posted above?

Well, I've styled the code element directly, but yes, this is it:

  border: 1px solid silver;
  border-radius: .3ex;
  padding: .1ex;

I'm apparently really bad at reading documentation, all along the ISO standard was linked for free on the GNU APL documentation page

though I don't know if I should add an 13751 dialect to the site, since apparently it's basically just apl2

hmm looks like another difference between the standard and apl2 is that the standard has lcm & gcd

@Adám better now?

@RubenVerg yes