« first day (2589 days earlier)   

10:39 AM
added apl\360 and apl.sv to omnibar
 
@RubenVerg It shows differences between SV and '89 but I'm pretty sure there are no such.
 
ah yea, I missed dyad | in the standard
 
Heh, missing ; for tuple in 360, SV and I think also in '89:
It isn't the same as Kap's tuple, but Omnibar doesn't care about such differences, does it?
 
it does if they're fundamentally different operations
what is the behavior or ; in 360/etc?
 
Exactly and only what you see in the screenshot (was it call a papershot in \360‽).
 
10:46 AM
@Adám the toc of iso '89 doesn't mention ;, doesn't necessarily mean that it doesn't have it, just that I can't find out
@Adám I don't understand what it's doing — my only guess is stringifying and concatenating, which sounds like a weird operation, especially because 360 didn't even have
 
That's exactly why it was needed. The old APLs didn't have heterogeneous arrays either, so how would you print a message like today's 'You have ',(⍕n),' items left'?
'You have ';n;' items left'
 
okay, in context i guess it makes sense
don't like "tuple" as a name though — is "catenate representations" fine?
 
No, it wasn't a function. You couldn't capture the result. It was purely for outputting multiple arrays adjacently.
 
i see. how would you call it?
tuple feels wrong because it's not creating a tuple
but I guess it's fine if we can't come up with anything better
 
Maybe "adjacent output"?
or "multiple output"?
 
10:53 AM
@Adám yeah, I like that more, thanks
 
or both: "multiple adjacent output"
outputs?
 
adjacent output is fine, no need to overthink it :)
 
Will you add .⎕⍞←¯E⍝ to SV or is there simply no reason to include SV since it is ≡ with '89?
 
no, as always I forget to do syntax, because it's not on the wiki
does 360 not have those?
 
It does have them all.
 
10:56 AM
also, are there quad names in SV or is ⎕ only I/O?
 
SV introduced quad names.
 
alright, should've added the missing entries
here's a question i had last night — should there be some order in the dialects? currently they're just sorted by order I added them in
I guess the options would be release date or popularity/influence?
though the latter is hard to judge and the former splits dyalog and ex/vi which feel like they belong together
maybe it's fine as it is, even if it's pretty much random order
there doesn't need to be a sorting
(also it's gonna be a chore to reorder all the entries! though I guess I could change the code so that I can specify the sort only once and then specify the dialects in whatever order)
 
11:24 AM
@RubenVerg I think release date makes sense, despite ex/vi.
@RubenVerg But is there really a reason to have both SV and '89 now?
Btw, a filter by/find glyph and meaning should really be a top priority now with almost 400 entries.
 
@Adám hmm, idk. they are different dialects even though they are equivalent
I think merging them would make the table less clera
 
You could use one tag for both, and just include them in the legend. There were loads of 360-equivalents too.
What about the OUT symbol? Doesn't exist in Unicode, afaik.
 
@Adám two search bars would be confusing, I think. is there anything wrong with using ctrl-f?
@Adám what's that?
 
@RubenVerg Yes, it isn't useful for common punctuation like .,/
 
@Adám I'd probably not include those. I think SV/the ISO standard are both important enough to deserve an entry, but it's an edge case
 
11:32 AM
@RubenVerg O<overstrike>U<overstrike>T was used to interrupt input.
 
@Adám right. not sure how to format the double search bar, though
unless i include syntax to search by glyph inside the same search but a) that's probably confusing and b) it requires a big change of how the query parsing and interpreting works
@Adám kinda weird, but alright. a) does it even belong in omnibar? b) what dialects had that?
also, what does "interrupt input" mean? is it just an alternative to hitting Return?
 
@RubenVerg I think the current one can be quite small. Nobody will make big set queries. And it should be called "filter". Then a larger bar on its left is actual search.
@RubenVerg No, it stops the program at that point.
@RubenVerg Certainly from APL\360 and until APL*PLUS.
 
was there an OUT button on keyboards?
 
Either that, yes, or you'd type it as I wrote above.
 
oh right, forgot that overstrikes were an actual thing you could do and not just a concept
@Adám so if OUT appears in a string handled by ⍞, the program stops?
 
11:39 AM
… which reminds me, you have APL+Win, but APLPLUS, APLPLUS II, and APL*PLUS III preceded it and each added something.
@RubenVerg Yes.
 
@Adám the usual question: is there (hopefully good) documentation available for those?
 
It exists… for the right amount of money!
 
@Adám I'm not sure what to classify it as, probably syntax?
 
But again, APL*PLUS is basically the same as APL.SV. APL*PLUS II is basically the same as APL2, and APL*PLUS III is basically the same as APL+Win.
 
@Adám ... of course.
 
11:42 AM
@RubenVerg Certainly nothing else.
 
@Adám are the apl*plus'es relevant enough to get double entries?
also probably won't add them if i can't have docs
 
I don't think so. But while APL*PLUS changed name (it had a version number in addition to the Roman numeral), Dyalog just incremented the version number, while adding significant features. Early Dyalog had e.g. no namespace . and was missing many primitives compared to today. Multiple entries?
 
yeah. i guess if i were to add apl*plus i'd add just the latest one, even if it changed name
but also, if they don't have free docs maybe they don't deserve to appear on the page :))
I understand making an interpreter paid, but the docs? those should always be freely available imo.
 
@RubenVerg Ah, OUT was added in APL.SV, so it wasn't in 360 (but it appears from there that 360 didn't have which it clearly did).
@RubenVerg Yes, that's insane. How do you expect people to buy your product if they have no idea what they're buying‽ (A: you don't expect people to buy your product!)
@RubenVerg Apparently, it was officially called "Heterogeneous output"
 
@Adám maybe they're scared of people stealing their new features!
 
11:51 AM
No, because all new features are described in freely available release notes.
@Adám Nope, misread. Its behaviour was slightly changed in SV, so it did exist in 360.
 
@Adám this document seems to imply was in \360, but the wiki doesn't have that (in fact, it explicitly says that SV introduced it)
oh, nevermind, I misunderstood the wiki
is there a list of the latest additions to \360? the wiki says it only lists the earliest version
hm, I guess if 360 evolved to be like SV maybe I could keep the earlier version on the table just so they're somewhat differenct
@Adám hm, I'll try that and see if it fits the page
it's already somewhat crammed on my small screen, adding another column (or two, with a label) would probably make everything too small
but I guess they can go on separate lines, if I clearly distinguish between filtering and searching
 
12:10 PM
@RubenVerg It isn't so much that \360 evolved to be like SV, but that SV was an extended (and slightly modified) \360.
Same with APL"1" (i.e. SV etc.) and APL2.
 
12:30 PM
Yeah so IBM implemented APL for a whole bunch of hardware, and mostly named it after what it was implemented on, or in some cases peripheral features. I think they just considered there to be one "APL" and updated the active stuff to match, not sure exactly. There's a detailed account of a few dozen systems in Adin Falkoff's article here (p. 416).
 
 
3 hours later…
3:02 PM
hm, I guess adding search for meaning names means I also have to document alternative names for the primitives?
time to look at all the source material for the third time I guess...
should the alternative names be displayed or just stored for search?
(this is somewhat annoying)
 
Just for search. That's what APLcart does too. Otherwise it becomes a clutter.
 
 
2 hours later…
4:40 PM
Announcement: Free DYNA24 event (NYC 11 April) for existing and potential users of Dyalog. Focus: migration and modernisation of APL systems.
 
 
4 hours later…

« first day (2589 days earlier)