A Localization Horror Story: It Could Happen to You • Hacker News Board

Re: A Localization Horror Story: It Could Happen to You

#251

Post by theoh » Sat, Jan 17, 2015, 3:17 AM UTC

lmm wrote: post #249

zokier wrote: post #246

Earlier quoted context omitted.

> there are plenty of other cases where they will change (e.g. precomposed characters that don't have a precomposed upper case). Just curious, is there still some cases if you only consider NFKD strings/characters?

Yes; the Turkish "I"s under discussion here are the most immediate case, but there are other cases where you have two almost-aliases in one case that aren't present in another case even ignoring composition. E.g. the ohm symbol "Ω" lowercases to a standard omega "ω", but that uppercases to a standard uppercase omega "Ω", because there's a distinct codepoint for "ohm symbol" (even though it's "just" omega, perhaps bec…

Is that actually the right thing to do, or is it another mistake?

Re: A Localization Horror Story: It Could Happen to You

#252

Post by davvid » Sat, Jan 17, 2015, 5:01 AM UTC

fennecfoxen wrote: post #127

TeMPOraL wrote: post #80

Earlier quoted context omitted.

I don't know why nobody seems to put information like "warning: this phone's UI in is total and utter crap". Anyway, what you wrote is exactly why I stick to using all software and webservices - OS, text editors, Facebook, et al. - in en_US instead of my native pl_PL. Because translations are always crappy - even for big players. Lack of context is the key here - translated text often feels out of place, because ther…

You know what multi-million movie has a translation that isn't total crap? Frozen . They really put resources into that. You can look up random Disney songs on Youtube in different languages, and then look up the Frozen songs, and you can sort of tell that they've done a better job even if you don't speak the language. Even relatively obscure languages like Dutch where they usually just watch English-language movies:…

For Big Hero 6 we also translated and replaced all of the Japanese text in San Fransokyo with Chinese and Korean for the Chinese and Korean markets. By "text" I mean all of the CG signs, posters, and environmental set dressing in the actual movie (not just the dialogue).

We literally had to re-render the entire movie for each translation. Disney Animation takes these translations very seriously :-)

Re: A Localization Horror Story: It Could Happen to You

#253

Post by Retra » Sat, Jan 17, 2015, 6:46 AM UTC

ufmace wrote: post #200

blywi wrote: post #111

Earlier quoted context omitted.

My favorite is still this one: http://www.theguardian.com/world/shortcuts/2012/feb/08/diese...

That's a head-smacker. Some guy somewhere must have actually painted that on the trailer - does that guy not think "Hey, this doesn't really make sense. Are you sure they didn't want somebody to translate this to Arabic?" Or maybe he did and his boss told him to shut up and do as he's told. Or he thought - "Screw this, I wanna go home on time. I'm not gonna bother to find an Arabic translator now. I'll just paint exa…

Maybe he had a stencil and didn't speak english?

Re: A Localization Horror Story: It Could Happen to You

#254

Post by zokier » Sat, Jan 17, 2015, 9:57 AM UTC

lmm wrote: post #249

zokier wrote: post #246

Earlier quoted context omitted.

> there are plenty of other cases where they will change (e.g. precomposed characters that don't have a precomposed upper case). Just curious, is there still some cases if you only consider NFKD strings/characters?

Yes; the Turkish "I"s under discussion here are the most immediate case, but there are other cases where you have two almost-aliases in one case that aren't present in another case even ignoring composition. E.g. the ohm symbol "Ω" lowercases to a standard omega "ω", but that uppercases to a standard uppercase omega "Ω", because there's a distinct codepoint for "ohm symbol" (even though it's "just" omega, perhaps bec…

> E.g. the ohm symbol "Ω" lowercases to a standard omega "ω", but that uppercases to a standard uppercase omega "Ω", because there's a distinct codepoint for "ohm symbol" (even though it's "just" omega, perhaps because some legacy codepages included it as a symbol without including a full greek alphabet) but no corresponding lowercase codepoint.

Except that the NFKD form (which I was specifically asking) for 'OHM SIGN' is 'GREEK CAPITAL LETTER OMEGA'.

Re: A Localization Horror Story: It Could Happen to You

#255

Post by lmm » Sat, Jan 17, 2015, 12:16 PM UTC

zokier wrote: post #254

lmm wrote: post #249

Earlier quoted context omitted.

Yes; the Turkish "I"s under discussion here are the most immediate case, but there are other cases where you have two almost-aliases in one case that aren't present in another case even ignoring composition. E.g. the ohm symbol "Ω" lowercases to a standard omega "ω", but that uppercases to a standard uppercase omega "Ω", because there's a distinct codepoint for "ohm symbol" (even though it's "just" omega, perhaps bec…

> E.g. the ohm symbol "Ω" lowercases to a standard omega "ω", but that uppercases to a standard uppercase omega "Ω", because there's a distinct codepoint for "ohm symbol" (even though it's "just" omega, perhaps because some legacy codepages included it as a symbol without including a full greek alphabet) but no corresponding lowercase codepoint. Except that the NFKD form (which I was specifically asking) for 'OHM SIG…

Ah sorry, I was sure I'd read NFD. Will look more.

Re: A Localization Horror Story: It Could Happen to You

#256

Post by lmm » Sat, Jan 17, 2015, 12:22 PM UTC

theoh wrote: post #251

lmm wrote: post #249

Earlier quoted context omitted.

Yes; the Turkish "I"s under discussion here are the most immediate case, but there are other cases where you have two almost-aliases in one case that aren't present in another case even ignoring composition. E.g. the ohm symbol "Ω" lowercases to a standard omega "ω", but that uppercases to a standard uppercase omega "Ω", because there's a distinct codepoint for "ohm symbol" (even though it's "just" omega, perhaps bec…

Is that actually the right thing to do, or is it another mistake?

As you said and linked elsewhere in the thread, the unicode consortium takes the viewpoint that there should be one codepoint for each glyph, even if that glyph has multiple semantic meanings in different languages (e.g. "U"). So by that standard they should probably be the same codepoint, but in that case it's hard to argue that roman capital I and turkish capital dotless I should be different codepoints.

Alternately you could argue that ohm symbol shouldn't lowercase to omega, which, maybe. I think the right view is simply that lower- and upper-casing aren't always well defined, are culturally and contextually dependent, and are probably something you should only ever be doing for display, not for semantic purposes. (If you want to do case-insensitive comparisons of strings, Unicode comes with algorithms for that which do a better job than upper- or lower-casing the strings before comparing)

Re: A Localization Horror Story: It Could Happen to You

#257

Post by hpshelton » Sun, Jan 25, 2015, 12:15 AM UTC

reidrac wrote: post #7

Slightly OT, mi favourite localization error was in Ubuntu when they had that nice netbook interface (that later would become Unity). The network icon label was "Rojo" in the Spanish localization, that is the word for "red" color. What? Well, if you translate "Net" to Spanish you get "Red"; and if you translate that again (by mistake), you get "Rojo". There you are :)

When I worked on Outlook.com, we got feedback from a British user that he couldn't understand what the option to "Connect devices and apps with DAD" meant. Turns out we had forgotten to prevent localization of the "POP" protocol everywhere it was used.