Hacker News Reader

Comment by veave | Hacker News Reader

Using software in a language other than US English has always been a pain.

Yup. Every graphic designer I know around the world uses Photoshop in English, not in their native language.

Every Photoshop user knows the difference between the image and the canvas, between layers and channels and levels. They learn what dodge and burn mean.

But nobody can even guess what arbitrary terms a translator will use for layers and canvases and channels and levels, or what terms printers used for dodging and burning back in the day.

Not to mention most Photoshop tutorials you find are in English as well.

And to someone who hasn't done touch-up work on film, the terms "dodge" and "burn" have no obvious connection to lightening or darkening an area, so even native English speakers need to learn what the terms mean.

Burn should be fairly logical. Burning things tends to make them black.

Russian translations tend to be quite robust even for open source software - large user base who are used to translation being there and of acceptable quality.

MS was the first mover here by having a robust translation of Windows 95 from the start (and yeah, they did translate the start button).

Russian seems to be better than even English at grabbing words wholesale from elsewhere. Maybe that makes translations easier? Probably not, but it's a fun thought.

My favorite example of that was a sign at an airport. Marking a spot off for taxis and rideshares. The latter was spelled out phonetically in Cyrillic. I was a little surprised they didn't use the Russian words for "ride" and "share".

A big, and already mentioned in this thread, issue is when ignorant people start to reimport words from other languages (English these days), and that catches up to some extent, even though educated people are using some proper term for decades or centuries.

You may end up having multiple conflicting ways to describe the same thing or even a person. A few days ago I saw Hades transliterated as Гадеc in some web comic translation, even though he is usually called Аид for many centuries already.

Open Source translations tend to be better in general. Specialy for smaller languages like Estonian. MS uses pretty horrible direct machine translations here that usually don't make sense.

Most people aren’t proficient enough in English. And even if they are, many would prefer to see their native language, just like you would prefer to see a nice GUI instead of MS-DOS.

(Though I personally find most Polish translations of software abysmal, especially the translations into Microsoft Polish, and use US/UK English everywhere.)

Microsoft's German translations at least used to be really good and professional. It's painful to see the decline.

Gotta convince the corporate Germans to switch to English somehow ;)

I haven't worked with MS products in ages, especially not in German. But I sure remember the gruesome "Schaltfläche" which unfortunately became the industry's standard translation for "button". And please don't forget about the "Eingabegebietsschemaleiste".

Professional? Maybe. Elegant? Not the least bit.

Since when?

Since forever according to the OP

Misspellings, mistranslations, strings that don't fit in the label/button, harder to google errors in other languages, etc.

What GUI toolkit are you using in 2023 where a widget doesn't adjust its size the size of its text label?

Literally any fixed-size dialog box ever.

Which seems to still be mostly the norm for apps in any desktop OS for things like preferences/settings dialogs.

Alot of UIs just can't deal with the size of a box being resized, so they instead adjust the size of the text, but for languages like Arabic and Japanese this can easily make text unreadable because characters often rely on finer details to be understood

There are so many other ways this can happen that it's really not surprising this happens. Like: started 20 years ago with the UI toolkit in English and a junior dev hence fixed size then suddenly got a growing customerbase in another language and slammed on a translation only to realize it would be a pretty large effort to track down and change each label to resize. Or even just: things adjusting size can simply be a big no depending on software (and you'd use truncation+toltip), because one doesn't want to waste space and/or keeps things aligned so it's easier to learn where everything is. Your browser tabs probably don't adjust size to text, just to name one thing.

Also text that's the wrong way or not ligaturized

These are localization issues. If the GUI was done in another language then US from start it wouldnt be an issue. Of course sloppy work is bad, doesnt mean it’s inherently bad

I was going to say since the invention of ASCII in 1963, but there were encodings before it and arguably you could say since the invention of Morse code in 1844.

The worst localization bug I'm aware of is the default conversion of string to datetime in MS SQL Server.

If the session language is non-US, the ISO formatted string '1991-02-03' becomes 2rd March rather than 3rd Feb. yyyyMMdd works as expected.

Reply on news.ycombinator.com