|
eschaton posted:yes, because it makes it at least theoretically possible to extract the development-localization string templates from the source code without having to parse (if not implement the semantics of) the full language I specifically meant from a "misuse of the operator" perspective, not i18n but as eschaton said using raw c string formatting has its pitfalls for i18n as well Private Speech fucked around with this message at 06:01 on Oct 14, 2020 |
# ? Oct 14, 2020 05:53 |
|
|
# ? Jun 13, 2024 05:34 |
|
I give absolutely zero fucks if I see "copying 1 file(s)", "copying 1 file", or "copying 1 files".
|
# ? Oct 14, 2020 07:31 |
|
You may not care, but the CEO will the next time (maybe next month?) he actually tried using the product
|
# ? Oct 14, 2020 08:00 |
|
Phobeste posted:the second, that’s pretty much boosts hallmark. I don’t think it’s apparent from the code but iirc in addition to overloading operator() (call) it’s overloading operator, (statement separator) for some reason
|
# ? Oct 14, 2020 08:29 |
|
Xarn posted:I give absolutely zero fucks if I see "copying 1 file(s)", "copying 1 file", or "copying 1 files". the extended version of which is "i give zero fucks if my software is user-hostile"
|
# ? Oct 14, 2020 08:44 |
|
Cybernetic Vermin posted:the extended version of which is "i give zero fucks if my software is user-hostile" ah, I see you too have heard of the unix philosophy
|
# ? Oct 14, 2020 08:44 |
|
Xarn posted:I give absolutely zero fucks if I see "copying 1 file(s)", "copying 1 file", or "copying 1 files". msdos user spotted
|
# ? Oct 14, 2020 10:07 |
|
I am not a native English speaker and I prefer by a factor of one billion percent that the software I use is not localised. I am still trying to figure out what improperly set bit in my environment makes GCC sometimes produce error messages partially in Danish (all the usual suspects are en_US.UTF-8, so I really have no idea).
|
# ? Oct 14, 2020 10:39 |
|
Athas posted:I am not a native English speaker and I prefer by a factor of one billion percent that the software I use is not localised. why? (I'm a native english speaker but I've come across/can imagine reasons for such a strong preference. so I'm not completely ignorant, just curious!)
|
# ? Oct 14, 2020 10:56 |
|
Athas posted:I am not a native English speaker and I prefer by a factor of one billion percent that the software I use is not localised. error messages and open source are both cases where localization is very bad (in slightly different senses) yeah, but properly done localization for everyday stuff is for the vast majority of people good and cool
|
# ? Oct 14, 2020 11:04 |
|
Athas posted:I am not a native English speaker and I prefer by a factor of one billion percent that the software I use is not localised. Good news! In my experience every localisation venture gets poo poo-canned as soon as the boss learns that it will cost more than an person-afternoon piping the code base through google translate. Plunk. straight into the shitter next to all the accessibility proposals.
|
# ? Oct 14, 2020 11:11 |
|
Athas posted:I am not a native English speaker and I prefer by a factor of one billion percent that the software I use is not localised. lmao i hate that also the locale env vars are a mess cause software will randomly use LC_ALL or LANG or whatever the gently caress with no rhyme or reason
|
# ? Oct 14, 2020 11:17 |
|
pokeyman posted:why? one reason: when you gotta google an unhelpful error message, you don’t want to have to start by trying to guess what the original English was before some volunteer dunning-krugered it into gibberish
|
# ? Oct 14, 2020 11:19 |
|
Cybernetic Vermin posted:the extended version of which is "i give zero fucks if my software is user-hostile" I prefer using sw with slightly broken grammar to "reticulating the splines" and similar poo poo. I would also willingly trade-off broken grammar for accurate progress bars, or CRC checks, or proper impl of atomic replacements, or ...
|
# ? Oct 14, 2020 11:32 |
|
Also thisAthas posted:I am not a native English speaker and I prefer by a factor of one billion percent that the software I use is not localised. Soricidus posted:one reason: when you gotta google an unhelpful error message, you don’t want to have to start by trying to guess what the original English was before some volunteer dunning-krugered it into gibberish Doing helpdesk by phone for my 80+ grandparents on their localized version of Windows is hell.
|
# ? Oct 14, 2020 11:54 |
|
Soricidus posted:one reason: when you gotta google an unhelpful error message, you don’t want to have to start by trying to guess what the original English was before some volunteer dunning-krugered it into gibberish yeah big time
|
# ? Oct 14, 2020 11:58 |
|
Xarn posted:I prefer using sw with slightly broken grammar to "reticulating the splines" and similar poo poo. well, i'll certainly trust that you'll put the time you save making comprehensive localisation impossible towards solving those entirely tractable goals.
|
# ? Oct 14, 2020 12:18 |
|
If you think that "write to a different file, then atomically rename it to the one you should replace" is intractable, we should continue the conversation here https://forums.somethingawful.com/showthread.php?noseen=0&threadid=3863535&perpage=40&pagenumber=529
|
# ? Oct 14, 2020 13:31 |
|
I’d definitely rather have bad grammar than ever see another error message that begins “oops”
|
# ? Oct 14, 2020 16:14 |
|
Xarn posted:accurate progress bars no such thing
|
# ? Oct 14, 2020 16:17 |
|
Soricidus posted:I’d definitely rather have bad grammar than ever see another error message that begins “oops”
|
# ? Oct 14, 2020 16:23 |
|
Athas posted:I am not a native English speaker and I prefer by a factor of one billion percent that the software I use is not localised. okay but you can read and write english. preferring the original language when it's one you're fluent in is not exactly a hot take. Xarn posted:If you think that "write to a different file, then atomically rename it to the one you should replace" is intractable, we should continue the conversation here that is significantly less atomic and foolproof than the posix manpages would lead you to believe. a naive "write new file, rename new file to old name" followed by a power failure can lead to the name pointing to the new file but none of the data you wrote being in that file. you need to actually wait for the data to be written to disk before the rename. you might think that you just need to call fsync() in between, but fsync() doesn't actually reliably report errors and fsync() reporting success could actually just mean that your write was discarded.
|
# ? Oct 14, 2020 16:35 |
|
Soricidus posted:I’d definitely rather have bad grammar than ever see another error message that begins “oops” got it changing errors to start with uwups
|
# ? Oct 14, 2020 16:36 |
|
ewps we did a fucky wucky in our code. aah! its just too hawd
|
# ? Oct 14, 2020 16:37 |
|
Xarn posted:If you think that "write to a different file, then atomically rename it to the one you should replace" is intractable, we should continue the conversation here if you think it is tractable, then i recommend https://danluu.com/deconstruct-files/
|
# ? Oct 14, 2020 16:37 |
|
Yeah, I expected that someone will make me die on that hill as soon as I posted
|
# ? Oct 14, 2020 16:40 |
|
Speaking of annoying progress """bars""", I had to use Webex yesterday. Apparently someone thought that "bringing in the chairs" is a good progress message for online teleconf. software.
|
# ? Oct 14, 2020 16:45 |
|
Bloody posted:if you think it is tractable, then i recommend https://danluu.com/deconstruct-files/ people who think their stuff works in these situations tend to not have spent much effort testing in these situations
|
# ? Oct 14, 2020 17:03 |
|
Volte posted:by totally fine i mean it's still a perfectly localizable string. it's still a guillotinable offense to use the bitshift operator for that purpose. but more than that, don't use any operator to feed parameters into a format string one at a time, not even a dedicated one
|
# ? Oct 14, 2020 17:04 |
|
and while it's hard to write files in the perfect way, doing write-then-rename still solves most of your problems and there's no excuse for not doing at least that much
|
# ? Oct 14, 2020 17:09 |
|
i used to think "oops" error messages were the worst, but i recently got one that started with "hmm..."
|
# ? Oct 14, 2020 17:13 |
|
also i've said this a bunch but it makes me lol that all voice assistants are essentially CLIs but somehow even less usable. like if you gently caress up command line syntax or use the wrong flags or whatever you know immediately what you did wrong when the terminal dumps out a wall of text, but if you say something the wrong way to alexa you're forced to wait like 5 full seconds for her to say "oh dear i'm sorry" at best, or at worst she says "ok... i found the following results for 'something you definitely did not ask for'"
|
# ? Oct 14, 2020 17:16 |
|
Plorkyeran posted:okay but you can read and write english. preferring the original language when it's one you're fluent in is not exactly a hot take. I'm not self-centered enough to think my own preferences are The Only Way To Do Software, but most of the software I use day-to-day is command like stuff where localisation is completely pointless. There was never a time in my life where I wanted my native language in the command line, even when I was a literal childe. Glibc and the coreutils have a hilarious amount of code in them to support dubious localisation functionality that should be done at a way higher level.
|
# ? Oct 14, 2020 18:04 |
|
Subjunctive posted:people who think their stuff works in these situations tend to not have spent much effort testing in these situations
|
# ? Oct 14, 2020 18:36 |
|
Athas posted:I'm not self-centered enough to think my own preferences are The Only Way To Do Software, but most of the software I use day-to-day is command like stuff where localisation is completely pointless. There was never a time in my life where I wanted my native language in the command line, even when I was a literal childe. Glibc and the coreutils have a hilarious amount of code in them to support dubious localisation functionality that should be done at a way higher level. okay, but again, you can read and write english. are you aware that there are people who use computers (and yes, even developer tools) who often have to resort to jamming poo poo into google translate?
|
# ? Oct 14, 2020 18:40 |
|
then they should go all the way and localize the c keywords and the stdlibcode:
|
# ? Oct 14, 2020 19:02 |
|
http://www.catb.org/esr/faqs/hacker-howto.html#skills4
|
# ? Oct 14, 2020 19:29 |
|
tw: esr
|
# ? Oct 14, 2020 19:31 |
|
Localization is good. I just don't give a poo poo about proper pluralization and similar details. But please, please, please, have your errors also emit a unique ID that people can google, so that they don't have to guess on what the original error message was.
|
# ? Oct 14, 2020 19:55 |
|
|
# ? Jun 13, 2024 05:34 |
|
ORA-12154
|
# ? Oct 14, 2020 19:59 |