The APL Orchard

@font-face {
font-family: 'savx2';
src: url('./fonts/Apl385.woff') format('woff'), url('./fonts/Apl385.ttf') format('truetype');
src: local('SAX2'), url('./fonts/Apl385.woff') format('woff'), url('./fonts/Apl385.ttf') format('truetype');
}

here is my ride-base.css file,

I add the last line to original file

on my RIDE 4.5, I use the name SAX2, it seems to me does not us the font

Adám

@1234 That's not what the instructions say to do.

@1234 Did you install SAX2 as you'd install any other font?

1234

I download the SAX2.ttf in the same folder as all other fonts

in APL wiki, it does mean replace the line?

In APL wiki, use the word 'insert'

Adám

@1234 I don't know what kind of system you have, but maybe there's something you need to do to actually "install" the font. You might also need to restart (OS or at least RIDE) to make the change be registered.

@1234 It means Insert the characters local('font name'), into the existing line, immediately after src:

1234

8:11 AM

do I need to keep the line?

src: url('./fonts/Apl385.woff') format('woff'), url('./fonts/Apl385.ttf') format('truetype');

I'm on macOS

Adám

So, yes, keep the line, but edit such that it says src: local('font name'), url('./fonts/Apl385.woff') format('woff'), url('./fonts/Apl385.ttf') format('truetype');

@1234 Can you use the SAX2 font from other applications?

@1234 Double-click the font in Finder, and Font Book opens and shows you a preview. Click "Install" and you're all set:

1234

work, :)

I thought I just download it and dump it in the font folder..

I do not know I need to install it...thx

RubenVerg

8:51 AM

how does the APLcart embed on the wiki work? it looks like it includes a PNG but then it's magically interactive?

Adám

@RubenVerg yeah, it is a bit of a hack, but seems to work very well. It is implemented in client-side JS.

I especially like the APLcart solution (as opposed to those of TryAPL and ngn/apl) because it provides an image fallback if JS is disabled.

B. Wilson

@Adám How do 64-bit integers work in APLX? Auto-promotion with the moral equivalent of some new ⎕DR class? Or is it more explicit?

RubenVerg

what dialect introduced flat with boxes?

(and what dialect introduced nested? the wiki seems to imply it's NARS)

Adám

@RubenVerg SHARP APL.

@RubenVerg Yes, but watch out; this is NARS, not NARS2000.

But before any implementation, there were years of presentations, papers, discussions…

@B.Wilson Any expressions you want me to run for you?

B. Wilson

@Adám I just found out that APLX "has 64-bit integers", so I'm naïvely wondering how that fits into the language.

Adám

9:08 AM

@B.Wilson After some experiments, I'd almost venture to say it doesn't:

      ⎕DR i←1234567890123456789
2
      ⎕DR f←2×1234567890123456789÷2
3
      i=f
1
      i=⌊f
0
      i=⌈f
0
      i-f
0
      i-⌈f
21
      i-⌊f
21

RubenVerg

I'm noticing most differences between dialects are historical — is it interesting to mention flat vs nested, when every current dialect is nested? is it interesting to mention LCM/GCD, when all dialects now support them? the only difference I can come up with that's actually a thing between current dialects is ⊃/↑ being swapped

Adám

APL2 doesn't extend ∧ (well, ^) and ∨ to LCM and GCD.

B. Wilson

@Adám Okay, so you'd expect that "has 64-bit integers" means auto-promotion and not explicit syntax? Here's my reference: github.com/Dyalog/aplx/blob/master/Differences.md

RubenVerg

@Adám oh, right

also real/complex is potentially a difference, I guess? it's not that interesting though

Adám

Right. APLX and APL+ have no complex, and thus limited ○ as well. Interestingly, ¯8*÷3 gives a domain error in APLX. I guess they didn't want to give a different result from APL2.

@RubenVerg ≡ in non-Dyalog APLs is |≡ in Dyalog.

RubenVerg

9:14 AM

all the sections of the article are like two sentences long, I really don't know how to properly format it

Adám

Don't worry, someone else can do that.

RubenVerg

9:27 AM

alright, created a stub for comparison of dialects

I'll let someone else add the omnibar embed since I don't understand how it works

I'll try and think of other significant differences and add them in the future

Adám

@RubenVerg Awesome. This is great raw material that others can build on. Exactly the type of collaborative work I was hoping for. Thank you!

RubenVerg

suggestions for the next dialect to add to omnibar?

@Adám love to help!

Adám

While officially dead, I think APLX and ngn/apl would be interesting.

Do you consider Kap an APL? April might be good too.

You could add A+ and SAX but there are quite some differences.

APL64 should obviously be added, but has to wait until I get access to its docs.

RubenVerg

I wanted to add Kap some days ago but gave up because there's too many differences :) it's definitely on the list though

Adám

APL+Win too.

RubenVerg

9:37 AM

I'll add APLX, looks like it's got good documentation

btw, any ideas for how to order glyphs? i remember you suggested by codepoint, but that seems somewhat wrong, some glyphs definitely belong together (</≤) which wouldn't

Adám

APLX should be fairly trivial too. Watch out for monadic ⍪ being different from Dyalog's. It has a whole unique primitive, ⍺. Then there are four quad-arrow file functions.

RubenVerg

currently the order is pretty much order-i-added-them-in plus some reordering for things that feel like they belong together (plus syntax at the end)

@Adám I assume ⍺ being a primitive implies no dfns?

Adám

@RubenVerg APLcart orders them roughly by conceptual complexity.

@RubenVerg Indeed.

RubenVerg

@Adám monadic ⍪? can't find that in the docs

Adám

It is well hidden. They define ⍪ as ,[1]

(Not entirely precise, as it also ravels scalars, but…)

RubenVerg

9:44 AM

I see, do they give it a name? (if not, suggestions for one?)

Adám

No, they don't. It is a completely useless function, as it is equivalent to monadic , — so same name as monadic , I guess.

(What a waste!)

@RubenVerg There's a difference in binding strengths: In APL2, array-array is weaker than indexee-bracket and operator-array — but not array-operator!

In Dyalog, array-array is almost the strongest.

RubenVerg

@Adám I'm not sure I understand what ⍺ does — could you provide an example?

it sounds similar to format by example

oh there's examples I just didn't read far enough, nevermid

Adám

It does. There are plenty of examples in the docs. But do you care for Omnibar? Just call it "Picture Format" and link to the doc.

RubenVerg

oh, is there a doc website? i'm looking at the language reference pdf on the dyalog page

Adám

Yeah, they have a really good online help: microapl.com/apl_help — but go here for primitives.

RubenVerg

9:58 AM

thanks

Adám

12:25 PM

@Marshall Would it make sense to merge Timeline of array languages and Timeline of influential array languages? — It sounds like the former is a superset of the latter, but the latter describes what's influential. Couldn't we just have an influence column, and maybe bold the influential events?

RubenVerg

1:03 PM

does APLX not have monadic ⊖?

Adám

It does. But again, same issue as with ⍪ that it is implied (under the dyadic docs!) that it is just ⌽[1] (except for scalars).

@RubenVerg Do you think you could compress the introductory parts of the Omnibar page a bit? This is how it will look on the wiki:

@RubenVerg I suggest:
1. Moving the GitHub link to be superscript (and/or right-justified) on the h1 title line.

2. Combining the lede and the bullet points into a single paragraph:
This tool allows you to compare the 320 glyph meanings of all (?) APL dialects, making it the ultimate language toolbar. Many dialects call the same primitive different names. In this list they are all grouped under a common name.

3. Make the two bulleted lists below the rainbow be inline lists:
x, y, u, v: Arrays m, n: Monadic functions d, e: Dyadic…

Btw, I think the 3-letter codes can be abbreviated to 2: ExD→DE, DVi→DV, N2k→N2, GNU→GA.

Marshall

1:21 PM

@Adám The full table is over twice as long. I don't think bolding or added explanation can really substitute for a condensed summary. I use both tables at various times and having them together just sounds like a lot of extra scrolling.

Adám

Fair enough, but then maybe we can improve their titles to make it clearer what they are?

RubenVerg

do I correctly understand the aplx docs that replicate is an operator?

@Adám sure

Marshall

What do you suggest? I don't find it confusing for a page with a narrower scope to have more detail; that's pretty much always the case.

Adám

@RubenVerg Yes, it is mostly a clone of APL2.

@Marshall OK, maybe it is fine then. I was looking at Template:APL dialects and thought it a bit confusing to see "Timeline of array languages ∙ Timeline of influential array languages"

Maybe it could (just in the navbox) be styled as "Timeline of array languages (Influences)" or something like that.

Marshall

I still think of the main purpose of the influential array languages as listing what they are rather than the added descriptions. And it doesn't seem strange for someone to look at those two names and think "no, I don't want to see every language, just the influential ones".

Adám

1:32 PM

Why would one want to know what they are but not what they contributed? For me, it was certainly mostly interesting to see when various features arrived, not so much who added them.

Marshall

Yes, but you already know what SHARP and APL2 and J are. For most people the existence of all these historical APLs is going to be new information.

Adám

Good point.

RubenVerg

better?

Adám

Yes, much. (I still find it strange that it says Omnibar twice above each other, but whatever.)

Shouldn't "Omnibar: the ultimate language toolbar" be the page <title>?

RubenVerg

idk, are <title>s supposed to be that long?

Adám

1:43 PM

Yes, they tend to either be "[sitename]: [short description]" or "[subpage] - [sitename]"

RubenVerg

pushed the changes

Marshall

The tables could be put on the same page. I don't think it's that unusual for a Wikipedia timeline to have a summary section and a complete list (also looks like articles are often titled "list of..." even if they're ordered by date).

Adám

I'd like that.

Marshall

Wikipedia has a separate list of programming languages and timeline of programming languages that includes most but not all of them (also some earlier timeline entries like PACT I aren't on the list). So at least we're doing better than that.

Adám

@RubenVerg I actually think you can make the dialects list even more compact (which will work much better on small screens):

RubenVerg

1:54 PM

@Adám I didn't do that because it didn't fit on one line on my screen, so thought it might look better formatted like i did

Adám

This is how it looks on the wiki in a narrow window with 150% zoom:

RubenVerg

anyways, added APLX, as usual I probably missed a lot of things

Adám

Looks alright, imo, as long as you don't line-break inside any badge-name pair.

RubenVerg

think I found a way to not have line breaks in the middle of the pairs, pushed it

Adám

Yes, that works. Nice.

@RubenVerg The rainbow and the paragraph following it are laid out such that it is less-than-obvious that the colours actually match, plus you have duplicated text. How about placing each sequence of letters centred below its corresponding colour rectangle and removing the adjacent text. Add a little white space between the coloured boxes too.

RubenVerg

2:05 PM

hmm, currently the rainbow is autogenerated and the paragraph is hardcoded

I'd have to probably hardcode both, not sure I want to do that, too much code duplication

I think it's clear enough as it is, even with the repeated text

Adám

At the very least, you could use the same order, i.e. move Arrays to the end.

RubenVerg

hmm, the categories are stored in a js object, I think technically there's no guarantee on the order

you actually think it's not clear enough rn?

Adám

No, it is clear enough, just looks a bit messy when it could look really sleek: