Live data from Hacker News

Don't guess my language

vitonsky.net

271–280 of 392 posts

Re: Don't guess my language

#271
post #240

Earlier quoted context omitted.

My computer is set to English even though I'm German, and sometimes Youtube will treat me to this really uncanny machine voice with really weird phrases because it auto-translated some German video or advert. Lidl is worth it, ja!

Same here. My native tongue is German, I live in Switzerland, but my settings on all devices are English. I do this on purpose, because I find everything is more searchable. I don't even know any German terms for most technical things I might search or look for. So even if the automatic translations were good, which they aren't, this would be a non-feature. My browser already tells them what my preferred language is.…

Living in Zurich (German part of Switzerland for those who don't know), Windows 10 in English, the built-in Microsoft Store used to offer content in... French.

Now it's a mix of German and English, e.g. 1 heading is "Spiele-Bestseller", and the next is "Best selling apps". And prices displayed as "28,00 CHF" (correct would be to use the decimal point).

Like Van Halen's brown M&Ms, it just shows how sloppily this thing is programmed: https://www.snopes.com/fact-check/brown-out/

Re: Don't guess my language

#272
post #23

I also hate the youtube "feature" that translates the titles of videos to your setting's language. This is so annoying. I can understand English and don't need these automatic translations.

Do you have en-US or en-GB as an alternate, lower-priority language?

If an English variant is in your Accept-Lang: headers, I'd hope YT wouldn't auto-translate English titles.

The other thing that Google might properly use is account-specific language settings. But if they're using GeoIP as has been suggested, I agree they're doing it wrong.

Re: Don't guess my language

#273
post #248
post #204

Earlier quoted context omitted.

What a catastrophe. You punish the wrong person, and even worse, a channel owner will not even receive that signal! The vast majority of channel owners with English content is not aware what's going on. A friendly e-mail to the channel owner explaining the problem and asking to manually disable auto-translation is much more likely to achieve what you want. If you want to get rid of auto-translation on a systematic le…

So what you're saying: Please complain through proper channels and hope they accept your input? Or should he just keep using the signals he gets and immediately clean up his feed? I actually see this as a feature. YouTube recommends a lot of garbage. I suggested that they improved my feed but they implemented this signal instead. I use the exactly this method to weed out a lot of content I do not care for. You cannot…

As a fellow non English native speaker, I concur with all of the above. But if you only have time for one sentence:

> A poor translation is worse than no translation

Re: Don't guess my language

#274
post #63

Not accepting Accept-Language is one of my major pet peeves. What makes it worse is that many multilingual websites translate their language-switching buttons and the list of languages to the current language .... which is beyond fucking stupid and defeats the purpose. Wikipedia does this right. The button to switch languages is clear, using a universal multilanguage icon, and a list of languages (using the name of t…

> in alphabetical order Well, it's in an order, but I don't know about alphabetical. I clicked on today's English featured article and looked at the languages: "中文", "Italiano" are "suggested", then the remainder are grouped by geographic region, and aren't particularly alphabetical. They appear to be in groups which are still not alphabetical. Europe seems to have a Cyrillic group but "Қазақша" is shown after "Украї…

> Well, it's in an order, but I don't know about alphabetical. I clicked on today's English featured article and looked at the languages

This depends on whether you are viewing desktop site or mobile site. It also depends on if you have a non-default skin set in your preferences.

Seems like desktop (vector-2023) does the region thing.

Mobile does alphabetical by language name (i imagine codepoint order but i didnt check)

Some other skins are alphabetical by bcp47 code.

Re: Don't guess my language

#275
post #23

I also hate the youtube "feature" that translates the titles of videos to your setting's language. This is so annoying. I can understand English and don't need these automatic translations.

So much this. I suspect the idea that a person speaks more than one language is absent in US silicon valley. Else I can't explain why youtube only lets you set one language. Heck, even google allows you to configure all spoken languages in your account, the very same google account you use for youtube. Yet youtube ignores it and has its own settings.

I think the issue is not speaking more than one language, but not preferring your native language over the original content's language. This is a very American-English-centric view of the world, where content is made for your language and your demographic. Consumption from outside the US is the exception.

In the rest of the world, and especially in Europe, this is the norm, not the exception. On one hand there is the prevalence of US English media (hello hollywood), US english literature (esp. in tech); and on the other hand cross-consumption between EU countries is much more common.

Oh, and the content ends up being in English too, because that's how to reach many people. We don't want those to be translated, because we don't want a double translation. This is something that the US / Silicon Valley mind cannot comprehend.

Re: Don't guess my language

#276
post #130

Earlier quoted context omitted.

Yes, that’s confusing and probably hard to find a good balance. Someone speaking Greek or Czech may expect to find their language around E (Ελληνικά) or C (Čeština), but nope, on Wiki it’s all the way after Z.

Don’t special characters always go after the Latin alphabet? I think this is pretty common, and fairly expected behaviour. Of course nothing is perfect but I feel like the way Wikipedia handles it is consistent.

Wikipedia uses UCA sort order in categories (depending on which lang wikipedia you are reading). Most other lists just sort using unicode codepoint order (in NFC). So it depends, but yes, for generated lists other than categories ascii characters usually come first.

Re: Don't guess my language

#277

> Every browser sends an Accept-Language header. It tells you what language the user prefers, not based on location, not based on IP, based on their OS or browser config. And yes, users can tweak it if they care enough. This is also a broken assumption. First, Accept-Language is an ordered list, and most daily-multilingual people don't have an absolute order of preference, and more a topical list of preferences. If I…

> If I read an English news site that has a translated French version, it doesn't matter if I'm most proficient in French, I'll want the English version. The solution there is not to abandon the very useful Accept-Language header, but to never offer half-baked translations. Is your website multilingual? Fine, but only offer fully translated and verified translation sets. It's OK to have your UI and your own content t…

> Fine, but only offer fully translated and verified translation sets

No, I do not want the translation at all, unless I say so. Really. Even if it is the best translation in the world. I want and need to choose the language I will read or watching something in.

Re: Don't guess my language

#278
post #151

Earlier quoted context omitted.

Don’t special characters always go after the Latin alphabet? I think this is pretty common, and fairly expected behaviour. Of course nothing is perfect but I feel like the way Wikipedia handles it is consistent.

Not in the Czech alphabet: a, á, b, c, č, d, ď, e, é, ě, ... Also, we regard 'Ch' as its own letter. So yeah, try sorting alphabetically. I'll wait.

Digraphs like Ch are common in a lot of languages. Wikipedia supports that fine on category pages. E.g. https://cs.wikipedia.org/wiki/Kategorie:CHKO_%C5%A0umava

If you want to see bizarre sort rules, look up how french sorts accent characters.

Re: Don't guess my language

#279

Ideally a browser should ask for languages and country when being installed, in order to correctly fill the Accept-Language field. But that doesn't happen because it would introduce friction.

Why? Browsers already apply the setting from the operating system which is correct in almost 100% of cases, and if you need more options you can change it in the settings.

Re: Don't guess my language

#280
post #59

> Every browser sends an Accept-Language header. It tells you what language the user prefers, not based on location, not based on IP, based on their OS or browser config. And yes, users can tweak it if they care enough. This is also a broken assumption. First, Accept-Language is an ordered list, and most daily-multilingual people don't have an absolute order of preference, and more a topical list of preferences. If I…

> First, Accept-Language is an ordered list, and most daily-multilingual people don't have an absolute order of preference, and more a topical list of preferences. Technically Accept-Language allows you to specify a "quality value"/weight for each language... https://developer.mozilla.org/en-US/docs/Web/HTTP/Reference/... IMO there is a lot of improvements that can be made by both the browser and websites: Websites s…

No, please, don't force a language list. I don't have a "ranking of the languages I want to read". I want the original language of the content, with a quick button to switch.
Post reply on HN