A few hopefully relevant thoughts about translating docs<p>1) translating a piece of english UI will almost double the length of it. That's annoying.<p>2) documentation is hard to write and to maintain, and sometimes it's hard to read and use documentation. Creating multilingual documentation is even harder, and it's nearly impossible to have an entire coverage of the docs in all languages, which leads to incomplete or outdated/incorrect documentation - which, as the saying goes, is worse than none. Multilingual software inevitably makes the documentation situation of a software worse. Trying to apply language A docs to a language B interface is hell.<p>3) documentation alone isn't part of the story; forum threads, answered support questions and that kind of discussion-style paradocumentation is ranges in importance from helpful to just full-on essential and that kind of stuff is never <i>really</i> translated. Trying to get help from a mostly english community for a software translated is hell.<p>4) The sublanguage used in a given UI isn't written in a given tongue; really, it's written in a pidgin made of the original language of the application <i>plus</i> the slang of the domain of the application, which is usually tied to one, say, culture or spoken language in particular; very large software also typically creates its own sub-language, which separates itself from actual tongue the UI could say to have been written in. After all, any description of a what a software does, expressed in a language evolved for the interaction of human beings in the real, material world, <i>must</i> be a vague and fragile metaphor. Translating metaphors is nearly impossible. It's hell. It's poetry, but when poetry is not the goal, it's hell.<p>That also means that to read the documentation and UI of an application amounts to learning the sublanguage of that application, which as I explained is different from the actual human language it was written in. I can understand that "save" means "serialize the relevant subset of the application state to a file" without knowing anything about the "dictionary" meaning of the English word "save". It's not the same thing. "Save" could be replaced by "blub" as far as I'm concerned, but it's convenient for people who speak english to associate "serialize" to the concept of preserving, of keeping. But it's not as important as the consistency of the language developed to describe the concepts and operations of that application. A single "blub" is better than "Save" <i>and</i> "Sauvegarder". Duplicating the pidgin a software develops for itself makes using that software a special kind of hell.<p>Conclusion: it's easy to end up in situations where I have to use french software here in Quebec. Most of my friends have laptops and phones with french UIs. It makes using them a pain in the ass.<p>I absolutely, deeply despise translated UIs; translating a UI is almost systematically a counter-productive activity. It's hell.
While i appreciate that they're trying to do good, for people interested in communicating with people from other cultures translated technological terms have always been a massive bane. I live in germany and while germans generally have little patriotism they're fierce defenders of their language, which led to all sorts of domain-specific terms being translated across many different kinds of media. Be it movies, books, technical manuals, programming instructions.<p>And for all of them it's a common theme that someone who believes they know a thing, might even consider themselves an expert, they have trouble then branching out of their linguistic bubble. When talking to experts in their domain, from another language, they suddenly cannot communicate anymore, even if they speak the other language fluently, as they need to learn an entirely new vocabulary just for basic interaction.<p>And this is not just a theoretical situation. I've had problems when talking about books i read in my youth, because often important words or even names had been changed and the person i was talking with knew the same characters, concepts or ideas i knew but neither of us knew under what names the other knew them. I learned Perl on the internet, on my own, using exclusively the english source documents; when talking to german Perl programmers i've repeatedly had issues because they had learned it in german courses which used german CS terms. Instead of actually discussing business we spent half the time sorting out what to call a thing. I also translated MMO content professionally for a while. While our translations were great and allowed the germans a perfectly fluid native experience, we found that many germans tended to switch the game to english after a while, because they couldn't talk to other players since the item names, skill names and location names were all different.<p>There are some situations where it's fine to use a translated term. Example: Mouse. In german the word is written Maus and pronounced identically. In such a situation where the difference is minimal, a translatio can be fine.<p>However in all other situation i believe translation of domain-specific words does damage in the long term.
It's an interesting article and I can certainly get behind the sentiment of making software more accessible to non-english speakers. That said, if another language already has a word with the desired meaning, why not use it instead of inventing a new one?<p>This is a two-way street. Much of the english vocabulary comes from other languages: trek, pundit, taboo, gauntlet, moped, etc. And many languages use english loanwords for technical terms. Just try translating "internet" to other languages. Even Japanese is "インターネット" (intaanetto)[1].<p>1. <a href="https://translate.google.com/#en/ja/internet" rel="nofollow">https://translate.google.com/#en/ja/internet</a>
> <i>“windows” became “eyes”.</i><p>That is quite possibly because the word “window” literally is “wind-ow” or “wind-eye” — i.e. an eye, or opening, for the wind to blow through, which it did before windows had glass, and instead had shutters, like eyelids.
There's an obvious bias here. The piece ends with "localisation may keep small languages alive". Isn't that adding to the problem? We don't try to keep small programming languages alive (except a few diehards). Many small languages are the problem itself. They're not something to aim for.<p>I should add that I'm not complaining about making phones more accessible to more people. That's absolutely great. But why try to go further and use it as an excuse to drag out languages that the world would be better off without?
I'm amused that the article fails to note that as of 2012 Windows was available in 109 languages[1]...<p>[1] <a href="http://blogs.technet.com/b/terminology/archive/2012/02/21/14-new-languages-announced-for-windows-8.aspx" rel="nofollow">http://blogs.technet.com/b/terminology/archive/2012/02/21/14...</a>
It would be interesting to review the full list of creative translations:<p>“Crash” became hookii (a cow falling over but not dying)<p>“timeout” became a honaama (your fish has got away)<p>“Aspect ratio” became jeendondiral, a rebuke from elders when a fishing net is wrongly woven.
Maybe the translators' jobs would be easier if we avoided jargon in our user interfaces. Are there any jargon words or phrases that are common in UIs that we could do just as well without?