|
snype
|
#
?
Feb 2, 2024 07:01
|
|
|
|
| # ? Jun 3, 2024 20:40 |
|
|
pseudorandom name posted:e.g. KOI8-R contains both A and Д and you need to be able to losslessly roundtrip from KOI8-R to Unicode and then back to KOI8-R Д is a Cyrillic D. Cyrillic A looks identical to Latin and Greek A. At least in upper case. Printed Cyrillic uses small caps, actual lower case is only used in italics or handwriting. That character set also includes Latin characters as well of course because everything has to be a superset of ASCII.
|
|
#
?
Feb 2, 2024 07:11
|
|
|
damnit echi
|
|
#
?
Feb 2, 2024 07:17
|
|
|
Sapozhnik posted:Д is a Cyrillic D. Cyrillic A looks identical to Latin and Greek A. At least in upper case. Printed Cyrillic uses small caps, actual lower case is only used in italics or handwriting. i cдп гёдd гцssiдп дlгёдdу, dёspiтё пфт sтцdуiпg iт
|
|
#
?
Feb 2, 2024 07:30
|
|
|
VikingofRock posted:Filesystems/OSs should just give every file a UUID and then use that. Then, allow users to set the display names to whatever they want. Users wanna have 50 different files on their desktop, all called "temp.txt" or whatever? I say, let them! congratulations you’ve invented the Apple Lisa
|
|
#
?
Feb 2, 2024 08:16
|
|
|
Sapozhnik posted:Should have done Latin-Greek-Cyrillic unification as well while they were at it there’s a lot more than that, it’s really all Phoenician what we really need are invisible language tags in text so we can tell the specific language that should be used for sorting, comparison, etc. on a per-substring basis but we don’t just need language tags, we need dialect tags too, so we can tell whether to use Swedish, Norwegian, Danish, Icelandic, or original rules for the different Norse dialects at least it’ll improve text to speech pronunciation
|
|
#
?
Feb 2, 2024 08:24
|
|
|
pseudorandom name posted:its the same reason Unicode contains a bunch of precomposed characters even though you can make them using combining sequences ah but is ö an o plus dieresis or is it a completely distinct character? the answer(s) might surprise you!
|
|
#
?
Feb 2, 2024 08:26
|
|
|
fart simpson posted:snype the first good post this thread has seen in pages
|
|
#
?
Feb 2, 2024 08:44
|
|
|
eschaton posted:at least it’ll improve text to speech pronunciation how about invisible ligatures that annotate glyph clusters with phonemes
|
|
#
?
Feb 2, 2024 09:10
|
|
|
Internet Janitor posted:how about invisible ligatures that annotate glyph clusters with phonemes slight issue there with exotic languages that have different phonemes in different regions, such as *cough* english (or was that the joke?)
|
|
#
?
Feb 2, 2024 09:52
|
|
|
didn't Bruce Lee unify his fist with Han in 1973?
|
|
#
?
Feb 2, 2024 10:24
|
|
|
i had a han sandwich yesterday edit: meant ham. unicooooode 
|
|
#
?
Feb 2, 2024 10:39
|
|
|
eschaton posted:ah but is ö an o plus dieresis or is it a completely distinct character? should r + n and m be unified into a single rn character? or are they different?
|
|
#
?
Feb 2, 2024 10:53
|
|
|
Whew. I finally caught up and got current with this thread, all the way from p. 1721 way back in July 2022. I don't know what else to say right now other than: hi tef, I've missed you.
|
|
#
?
Feb 2, 2024 11:13
|
|
|
eschaton posted:there’s a lot more than that, it’s really all Phoenician we also need a separate aa digraph so Aabenraa sorts after Aachen and Zephyr Carthag Tuek fucked around with this message at 15:31 on Feb 2, 2024 |
|
#
?
Feb 2, 2024 15:28
|
|
|
eschaton posted:there’s a lot more than that, it’s really all Phoenician Speaking of sorting, I did a query against an SQL Server database where the value I was looking up was LIKE 'M%' and sorted by that value. The results came in with a bunch of 'M-A%' values through 'M-I%' values, followed by a handful of 'MJ%' values with no hyphen, followed by 'M-K%' and all the rest. I don't know what the hell collation settings managed to ignore hyphens while sorting, but I'm amazed they exist.
|
|
#
?
Feb 2, 2024 16:09
|
|
|
CPColin posted:Speaking of sorting, I did a query against an SQL Server database where the value I was looking up was LIKE 'M%' and sorted by that value. The results came in with a bunch of 'M-A%' values through 'M-I%' values, followed by a handful of 'MJ%' values with no hyphen, followed by 'M-K%' and all the rest. I don't know what the hell collation settings managed to ignore hyphens while sorting, but I'm amazed they exist. maybe they were soft-hyphens lol like ideally they shouldnt affect sorting and only show up when there's a linebreak there, but iirc ive seen them in terminal output (definitely as code points, which could maybe get rendered as regular hyphens if the settings are all hosed up?)
|
|
#
?
Feb 2, 2024 16:17
|
|
|
If they're soft hyphens then my predecessor was even more nuts than I previously thought 
|
|
#
?
Feb 2, 2024 16:37
|
|
|
CPColin posted:If they're soft hyphens then my predecessor was even more nuts than I previously thought lol they are 100% soft hyphens
|
|
#
?
Feb 2, 2024 17:01
|
|
|
CPColin posted:If they're soft hyphens then my predecessor was even more nuts than I previously thought i had a helluva time with them in a tool i wrote basically it was for copy-editing ocr, so what we did was split everything into tokens = "words". but since words are sometimes split with hyphens, we also checked for hard/soft hypens at end of token and marked them as hyphenated (to keep coordinates for editing the ocr in the pdf). but in the front end, we needed the editors to be able to see where the hyphens were, so it ended up being something like: - backend sends a "consolidated token" to the frontend, which had both left and right tokens, with the soft hyphen replaced by a hard one - frontend editor does whatever, maybe moves hyphens or something idk - backend then splits on the hard hyphen, turns it back into a soft, and saves the two tokens again just a real hacky job, but it actually worked, as long as therer werent actual hard hyphens in words that also included soft hyphens. we just skipped those lol
|
|
#
?
Feb 2, 2024 18:14
|
|
|
apple open sourced a DSL for configuration, because we don't have enough config formats https://pkl-lang.org/index.html it's a language that compiles to json/yaml/plist/properties and also generates static types for java/kotlin/swift/go and has plugins for intellij/vscode/neovim. i'm sure all of these things are supported equally and everything works consistently across the board
|
|
#
?
Feb 6, 2024 18:08
|
|
|
I wonder how it compares to cue
|
|
#
?
Feb 6, 2024 18:12
|
|
|
abraham linksys posted:apple open sourced a DSL for configuration, because we don't have enough config formats https://pkl-lang.org/index.html
|
|
#
?
Feb 6, 2024 18:12
|
|
|
pretty sure we have a ton of programmers using all of those languages and ides, so they’re probably all well-supported or are you under the impression that apple just doesn’t have a website
|
|
#
?
Feb 6, 2024 18:17
|
|
|
abraham linksys posted:apple open sourced a DSL for configuration, because we don't have enough config formats https://pkl-lang.org/index.html ah good, bindings for java, kotlin, swift, and go. you know, the languages people use
|
|
#
?
Feb 6, 2024 18:32
|
|
|
typescript would probably be a good target to add, i guess
|
|
#
?
Feb 6, 2024 18:57
|
|
|
I'm very much OK with dissing on Apple (give me OpenMP!!!), but I can't really begrudge some Apple project for only supporting code generation for four languages in its initial release. Since the thing is on GitHub, I don't think they would mind your contribution of support for Standard ML or other such superior languages.
Athas fucked around with this message at 19:15 on Feb 6, 2024 |
|
#
?
Feb 6, 2024 19:11
|
|
|
i don't think the point was so much the precise breadth of supports but a question of the commitment to the thing. which, i think it fair in that you probably should not expect *that* much more than from a random github repo with some momentum behind it. i.e. more of the modern "google releases a thing!" than ye olde "here's one of the building blocks of our technology stack, expect it deeply ingrained in our base sdk's".
|
|
#
?
Feb 6, 2024 19:15
|
|
|
i don't know the people behind pkl, so i can't directly vouch for them or tell you that it's already used in a million different places or anything like that, but i don't think apple has a particularly bad track record about creating and then abandoning developer products. i certainly don't think we've done anything to deserve to be named in the same sentence as google on that front
|
|
#
?
Feb 6, 2024 20:40
|
|
|
like i know that's just an open invitation for people to jump in and make a million complaints about xcode and lldb, but those products are clearly not actually abandoned, they actually get a lot of development, it's just that that development doesn't always fix whatever problems people are seeing (and also people like to bitch about their tools)
|
|
#
?
Feb 6, 2024 20:46
|
|
|
eh just point out cocoapods broke everyone yesterday and they'll take their xcode-shaped pitchforks elsewhere
|
|
#
?
Feb 6, 2024 20:48
|
|
|
rjmccall posted:but i don't think apple has a particularly bad track record about creating and then abandoning developer products. only people who don’t distinguish between the Apple of the mid-1990s and the Apple of today think this which isn’t to say things haven’t been deprecated and canceled and removed, but that’s platform evolution for you
|
|
#
?
Feb 6, 2024 21:26
|
|
|
rjmccall posted:like i know that's just an open invitation for people to jump in and make a million complaints about xcode and lldb, but those products are clearly not actually abandoned, they actually get a lot of development I know, apple should really stop trying to make dev tools. It's clearly not their strong suite.
|
|
#
?
Feb 6, 2024 22:42
|
|
|
You're right in that Apple doesn't have a reputation for abandonment like Google does. But stability, debugging experience, and cross-compatibility with other tools/ecosystem is what makes me want to avoid them.
|
|
#
?
Feb 6, 2024 22:59
|
|
|
you understand this is just like a build-time code generator right
|
|
#
?
Feb 6, 2024 23:16
|
|
|
This config file language has insufficient stability, debugging experience, and cross-compatibility with other tools/ecosystem. It probably only supports a single external monitor, too. And I bet it has disabled OpenMP! Edit: this would never have happened under Steve Jobs.
|
|
#
?
Feb 6, 2024 23:39
|
|
|
brb, filing a radar to add openmp support to tkl just for you
|
|
#
?
Feb 6, 2024 23:41
|
|
|
i wonder which is better: apple for people who work at apple, or google for people who work at google
|
|
#
?
Feb 7, 2024 00:52
|
|
|
DELETE CASCADE posted:i wonder which is better: apple for people who work at apple, or google for people who work at google everyone is too busy counting their figgies to care
|
|
#
?
Feb 7, 2024 01:11
|
|
|
|
| # ? Jun 3, 2024 20:40 |
|
|
mostly i just figure this is a relatively complex language to implement code generators for and keep them compatible across four languages: https://pkl-lang.org/main/current/language-reference/index.html and i don't really have any inherent reason to trust apple with open source other than, like, they do choose to open source way fewer things than other companies, so maybe that means they'll put slightly more effort into what they do release. i'm just inherently skeptical, got burned by the official google grpc/protobuf stuff having the worst loving typescript compilation imaginable and an entire ecosystem of third party tooling exists just to replace it, and protobuf is, like, 1/20th the complexity of this format
|
|
#
?
Feb 7, 2024 01:42
|
|