« first day (2435 days earlier)      last day (438 days later) » 

01:13
In the Dyalog desktop GUI application (on Mac, if relevant), is it possible to always show the backtick keyboard shortcuts below or next to the glyphs, instead of only on hover? Separate question: is it possible to enable the same shortcuts used in tryapl?
 
3 hours later…
04:04
@mitchelljohnstone Yes, of course. You can either tell me here or email me (adam@ with the same domain as dyalog.com) or PR.
@Jonah Always showing the shortcuts would explode the size of the language bar. How about printing out an overview instead? And by "the same shortcuts used in tryapl" do you mean the ASCII composition like -> becomes ? If so, no, but I'll take it up if you confirm.
04:27
@Adám My solution that passes on the online judge is ⊢≡⍥(1 1∘⍉+.,⊢)⌽∘⍉. Currently it doesn't really check that all of the sums are the same, so it should fail. A breaking case would be some kind of semi-magic square, such as ↑(1 0 0 1)(0 1 1 0)(0 1 1 0)(1 0 0 1), which should give 0, as all rows/columns sum to 2, but the diagonals both sum to 4.
Ah, I remember this problem. Thanks.
@Adám This problem? Had it come up before?
Yes, these are just old competition problems, and iirc, we hit this snag back when the competition was running.
@Adám Yes, I mean the stuff like "rr tab" for rho, etc. I found those intuitive. "would explode the size of the language bar. " What I am imagining would double its thickness, which would be a great tradeoff for me while learning. It wouldn't be the default, ofc, just something you could turn on. Better than printing out because its right there inline.
As an alternative for now, you can open the Preferences' Keyboard tab.
04:37
Yes that should work
 
2 hours later…
06:15
@mitchelljohnstone you'll see how actually beautiful Haskell is when I get into parsing (assuming I'll stick to using parsec like I planned to)
06:52
@RubenVerg Done.
@mitchelljohnstone Added.
07:14
@Adám thanks :)
 
5 hours later…
12:44
@Adám Thanks as well! :)
13:08
Man, I wish there were a way to define your own single glyph names so that abutting them together would tokenize them separately. Kind of feels like a blemish to have a random space amongst a bunch of primitives.
@B.Wilson Allowing assignment to glyphs is a request often made by those that have not taken the time to think of the implications :-)
@Adám I don't mean the squiggles. I just mean like z and I could be assigned so that zI⍳10 would tokenize the same as z I⍳10.
Ah, I see, but how exactly would that work? What happens to names (include system names!) that contain z or I?
System functions all have the marker to disambiguate, but dotnet stuff wouldn't play along well.
None of the system names use diacritic letters, right?
Might be able to have a restricted character set that just always become single-char tokens.
@B.Wilson None other than ⎕Á/⎕Ⓐ
13:23
Never heard of ⎕Ⓐ. Gives me a syntax error.
It is ⎕Á by you, but ⎕Ⓐ in some legacy versions.
Nifty.
@B.Wilson Hm, how about allowing these:
Mathematical Alphanumeric Symbols is a Unicode block comprising styled forms of Latin and Greek letters and decimal digits that enable mathematicians to denote different notions with different letter styles. The letters in various fonts often have specific, fixed meanings in particular areas of mathematics. By providing uniformity over numerous mathematical articles and books, these conventions help to read mathematical formulas. These also may be used to differentiate between concepts that share a letter in a single problem. Unicode now includes many such symbols (in the range U+1D400–U+1D7FF...
@Jonah Got any errors in the browser developer console (F12 on chrome/firefox)?
@Jonah @RikedyP I believe the help site doesn't serve the APL font it uses with @font-face, so if a user doesn't have the font installed, it'll look like this.
13:35
@Adám Ooooh. That's a delicious idea. Also allows for possibilities to annoy your friends: π←4.
@Adám Yeah just talked with Fi - her new project has @font-face but the current one doesn't right now
Is it really that hard to inject some CSS that fixes the problem?
 
3 hours later…
16:16
@Adám It could at least provide some alternatives, ie font-family: 'APL385 Unicode', 'APL386 Unicode', 'BQN386 Unicode', 'APL2741 Unicode', 'APL333', monospace; or something like that
can't remember the exact font name for most of these
@RubenVerg Hey, if we can inject CSS, then the problem would already be solved.
I would assume changing one line of CSS is easier than hosting a whole new file on the backend :)
I do have an userscript that injects .Dyalog, .DyalogExample, .DyalogPlain, .DyalogDefault, .APLCode, .APLCodeSmall, .APLCodeNoIndent { font-family: 'BQN386 Unicode' !important; }
But it could just get the font from somewhere else.
I have a userfont that injects div.message pre,div.message code{font-family:"APL385 Unicode"} to SE chat.
that looks overspecific - does it select eg the starboard code on the right?
I'm sure pre, code would work just as well
\○/ It works :-)
 
3 hours later…
19:49
@Adám Thanks Adám, installing the font worked. Two followups: 1. On mac, to install the font, I had to double click on the .ttf rather than just placing it in /Library/Fonts (just an FYI you may want to note on the dyalog website instructions. 2. Is it possible to just load the font using CSS like when you include google fonts? I also ask this selfishly in case I want to include APL in a blog post or something like that.
20:17
@Jonah 1. Forwarded to website person. 2. @font-face{font-family:'<font name>';src:url('<url to font>')} should work, e.g. @font-face{font-family:'APL385 Unicode';src:url('https://abrudz.github.io/APL386/APL386.ttf')}
20:48
@Adám Thanks.

« first day (2435 days earlier)      last day (438 days later) »