Revision history [back]

As to the basic idea of translating URLs I'd say yes, definitely translate them. It's a beautiful little detail.

It gets tough with the Unicode, though. While all modern browsers can, to my knowledge, deal effortlessly with Unicode characters in URLs, it is possible that a translated URL shows up in many other places:

• In text files
• Copy+pasted into E-Mails and other forms of communication
• As a link on an external web page that has different encoding

and so on. There is always the possibility that the encoding gets screwed up at some point. It's definitely worth an SO question, I (strangely!) can't see one dealing with this very issue yet. I'll see whether I'll whip one up later.

For latin-based languages, I would at the moment say translate them, but use ASCII characters only. Every language has their own, safe rules how to deal with that (in German, ä becomes ae, in Finnish, ä becomes a, and so on).

It's obviously not a solution for the rest of the world yet - I'll try to get more info.

As to the basic idea of translating URLs I'd say yes, definitely translate them. It's a beautiful little detail.

It gets tough with the Unicode, though. though, I think. While all modern browsers can, to my knowledge, deal effortlessly with Unicode UTF-8 characters in URLs, it is possible that a translated URL shows up in many other places:

• In text files
• Copy+pasted into E-Mails and other forms of communication
• As a link on an external web page that has different encoding

and so on. There is always the possibility that the encoding gets screwed up at some point. It's definitely worth an SO question, I (strangely!) can't see one dealing with this very issue yet. I'll see whether I'll whip one up later.

For latin-based languages, I would at the moment say translate them, but use ASCII characters only. Every language has their own, safe rules how to deal with that (in German, ä becomes ae, in Finnish, ä becomes a, and so on).

It's obviously not a solution for the rest of the world yet - I'll try to get more info.

As to the basic idea of translating URLs I'd say yes, definitely translate them. It's a beautiful little detail.

It gets tough with the Unicode, though, I think. While all modern browsers can, to my knowledge, deal effortlessly with UTF-8 characters in URLs, it is possible that a translated URL shows up in many other places:

• In text files
• Copy+pasted into E-Mails and other forms of communication
• As a link on an external web page that has different encoding
• Read by client libraries that can't deal with UTF-8 characters (see for example here)
• Viewed in non-standard browsers

and so on. There From what I can see looking around on SO, this is always the possibility that the encoding gets screwed up at some point. It's definitely worth an SO question, I (strangely!) can't see one dealing with this very issue yet. I'll see whether I'll whip one up later.possible, but shaky.

For latin-based languages, I would at the moment I'd say translate them, but use ASCII characters only. That's how I would do it in the german translation, for one. Every language has their own, safe rules how to deal with that (in German, ä becomes ae, in Finnish, ä becomes a, and so on).

It's obviously not a solution for the rest of the world yet - I'll try to get more info.

world, and it would be nice to be able to have cyrillic, chinese, japanese, korean...... question texts in the URL. Hmm.

As to the basic idea of translating URLs I'd say yes, definitely translate them. It's a beautiful little detail.

It gets tough with the Unicode, though, I think. While all modern browsers can, to my knowledge, deal effortlessly with UTF-8 characters in URLs, it is possible that a translated URL shows up in many other places:

• In text files
• Copy+pasted into E-Mails and other forms of communication
• As a link on an external web page that has different encoding
• Read by client libraries that can't deal with UTF-8 characters (see for example here)
• Viewed in non-standard browsers

and so on. From what I can see looking around on SO, this is possible, but shaky.

For latin-based languages, I'd say translate them, but use ASCII characters only. That's how I would do it in the german translation, for one. Every language has their own, safe rules how to deal with that (in German, ä becomes ae, in Finnish, ä becomes a, and so on).

It's obviously not a solution for the rest of the world, and it would be nice to be able to have cyrillic, chinese, japanese, korean...... question texts in the URL. Hmm.

Update: I've posted a question on SO.

As to the basic idea of translating URLs I'd say yes, definitely translate them. It's a beautiful little detail.

It gets tough with the Unicode, though, I think. While all modern browsers can, to my knowledge, deal effortlessly with UTF-8 characters in URLs, it is possible that a translated URL shows up in many other places:

• In text files
• Copy+pasted into E-Mails and other forms of communication
• As a link on an external web page that has different encoding
• Read by client libraries that can't deal with UTF-8 characters (see for example here)
• Viewed in non-standard browsers

and so on. From what I can see looking around on SO, this is possible, but shaky.

For latin-based languages, I'd say translate them, but use ASCII characters only. That's how I would do it in the german translation, for one. Every language has their own, safe rules how to deal with that (in German, ä becomes ae, in Finnish, ä becomes a, and so on).

It's obviously not a solution for the rest of the world, and it would be nice to be able to have cyrillic, chinese, japanese, korean...... question texts in the URL. Hmm.

Update: I've posted a question on SO.

with good results. Apparently, percent-encoded URLs will work fine, and displayed correctly in the browser. Evgeny, what do you think?