28

My real Romanian name is: Ionică Bizău. You can see ă letter that is very close to a (in both: writing and pronouncing).

After setting my username as Ionică Bizău the link to my profile became:

https://stackoverflow.com/users/1420197/ionic-bizu

I know that if I write

https://stackoverflow.com/users/1420197/ionica-bizau

I am redirected to my profile, but this confuses me every time I go to my profile seeing ionic-bizu in the link...

I would suggest to remove this or to replace the diacritics with normal letters.

e.g. For Romanian diacritics:

+-----+----+
| ă   | a  |
+-----+----+
| â   | a  |
+-----+----+
| î   | i  |
+-----+----+
| ș   | s  |
+-----+----+
| ț   | t  |
+-----+----+

Update: If keeping the url containing special characters is ok, it's very good. My problem is that the letters from my name are removed.

Update:

I just found that ë is replaced with e in the profile url. See Pëkka's profile. Its url is:

meta.stackoverflow.com/users/138112/pekka`

                                     ^--- Do you sëë? :-)

Temp solution

I finally decided to use ã character that is very close to ă (but not the same!). My SO profile url looks fine: https://stackoverflow.com/users/1420197/ionica-bizau

On the other profiles I still keep the old username, so I still want this to be fixed!

  • 6
    Wikipedia keeps actual diacritics in the URL without apparent problem. – TRiG is Timothy Richard Green Jan 1 '14 at 16:09
  • 5
    This varies greatly between languages, btw. Norwegian å is pronounced 'o' in other languages, for example, and the Norwegians replace it with aa when limited to ASCII characters. This is not always a straightforward transformation. – Martijn Pieters Jan 1 '14 at 16:09
  • @TRiGisTimothyRichardGreen That's a good point. If that's possible it would be great. My problem is that letters from my name are removed. – Ionică Bizău Jan 1 '14 at 16:10
  • 1
    The name part of your URL is just a courtesy, a human-readable addition. The number part before it is what counts. You are perfectly free to link to stackoverflow.com/users/1420197/ionica-bizau anywhere else, as you already discovered. – Martijn Pieters Jan 1 '14 at 16:10
  • 2
    @MartijnPieters I know that, but in my case it becomes a human-nonreadable addition because I see ionic-bizu every time when I go to my profile! – Ionică Bizău Jan 1 '14 at 16:13
  • 2
    Sure, I can see your point too; and mapping non-ASCII characters to ASCII characters based on decomposition is quite straightforward (å can also be expressed as the composition of a and the ` ̊` COMBINING RING ABOVE codepoint U+030A, and any good unicode data library will give you access to the composed form and create a simple algorithm to replace å with a. It all gets a little more complicated when it comes to alphabets not based on the western character set, of course. – Martijn Pieters Jan 1 '14 at 16:22
  • 1
    @MartijnPieters Yes... Keeping them would be better - if possible. – Ionică Bizău Jan 1 '14 at 16:24
  • 1
    In German, the correct transliteration of ü is ue; in Spanish, it is likely u. I'm sure there are more such localized distinctions – Pëkka Jan 1 '14 at 16:33
  • This is just some missing characters in the translation. See Jeff's answer in Non US-ASCII characters dropped from full (profile) URL. – Arjan Jan 1 '14 at 17:00
  • @Arjan Good point! ã is replaced, but ă not. Hope it will be fixed soon. – Ionică Bizău Jan 1 '14 at 17:03
  • I've retagged as bug, and reposted my comment to make it more visible. – Arjan Jan 1 '14 at 17:06
  • But why wouldn't you want to be ionic? Sounds pretty cool to me. – michaelb958 May 14 '14 at 1:30
5

We've had this a while for our "sites about languages", but we have now enabled it accross the board for all sites. Your name should now show correctly in your browser url, as should questions. For example:

http://meta.stackexchange.com/users/205508/ionic%C4%83-biz%C4%83u

which should display visibly as:

[http://]meta.stackexchange.com/users/205508/ionică-bizău

Note that we have not enabled non-ASCII tags for all sites at the same time - these are controlled separately.

  • 1
    Wow, thank you so much. It works fine! – Ionică Bizău Jun 10 '14 at 12:43
  • The issue is still present on http://stackexchange.com/ and http://data.stackexchange.com/. – Ionică Bizău Sep 30 '14 at 8:50
  • @IonicăBizău try SE.com again; I haven't looked at data.SE yet - that is a very different codebase – Marc Gravell Sep 30 '14 at 10:06
  • Fixed on SE. Thanks! – Ionică Bizău Sep 30 '14 at 10:14
  • @IonicăBizău and I didn't accidentally cause an infinite redirect loop for a few minutes there... ahem – Marc Gravell Sep 30 '14 at 10:15
7
+200

As TRiG mentions in the comments, the name could be retained completely using percent encoding:

http://stackoverflow.com/users/1420197/Ionic%3F+Biz%3Fu

It's a valid URL, and modern browsers will display the decoded version as a courtesy. (Using the Unicode characters themselves would result in an invalid URL.)

Seeing as the numerical ID is used to look up the record, and the real name is there for human eyes only, this shouldn't be a huge problem.

It would be more internationally friendly. A transition process could start changing the profile URL the next time the name is changed (for example).

  • Why percent encoding? Why not the character? Please see this site: http://www.scriuromânește.net/ – Ionică Bizău Jan 1 '14 at 16:30
  • 2
    @IonicăBizău the actual characters are never valid in URLs. Browsers may accept them and automatically percent encode them but it doesn't mean it's a valid practice. – Pëkka Jan 1 '14 at 16:32
  • But... Just asking: why your ë from username is replaced with e? :-) http://.../138112/pekka – Ionică Bizău Jan 1 '14 at 16:34
  • 1
    @Ionica hah, good point! I assume they have transliterations in place for some characters. In that case they really should add the Romanian ones (although simply showing all non-Latin characters would of course be the best option) – Pëkka Jan 1 '14 at 16:36
  • YES! :-) I wait for their answer. – Ionică Bizău Jan 1 '14 at 16:37
  • 1
    But would you prefer /p%3Fkka over /pekka...? My Chrome does not show a ë when hovering such link. – Arjan Jan 1 '14 at 17:08
  • @Arjan but it will in the address bar. It's arguably much fairer to anyone not using Latin script. – Pëkka Jan 1 '14 at 17:14
  • When typing (pasting) a percent-encoded URL in the URL bar, it stays encoded. So, encoding in <a href=...> would be fine (like here, with a fake id to make SE not rewrite it), but sharing it otherwise might show the percent encoding. (Like when copying and pasting into an email?) I don't know what I like better... – Arjan Jan 1 '14 at 17:28
  • (Hmmm, it doesn't even work in <a href=...> -- I'm sure it once did...? Or is that Markdown messing up? No, it's a proper <a href="http://meta.stackoverflow.com/users/-999/p%3Fkka">here</a> in my comment above.) – Arjan Jan 1 '14 at 17:29
  • 1
    @Arjan you need to use percent-Encoded UTF-8 characters. Example: P%C3%ABkka – Pëkka Jan 1 '14 at 17:33
  • Ah, of course. And now hovering looks fine too :-) – Arjan Jan 1 '14 at 17:36
  • Downvoted for looking ugly. No, not you. Even though you do wear a stallman beard. But slugs in URLs should usually contain just ASCII, and percent-encoded chars in there do look horrible. And when pasting the link in a place where URLs are linkified automatically chances are good non-ascii breaks it. I've seen too many broken links in email clients because of things like that... – ThiefMaster Jan 1 '14 at 17:43
  • 1
    @Thief but how do you transliterate 蒋中正? – Pëkka Jan 1 '14 at 18:08
  • 1
    @ThiefMaster yes, there are transliterations for foreign alphabets, but they're not trivial to implement, there are different transliteration systems for the same language, and they of course all assume that the West with its latin alphabet is the center of the world around which everyone else is supposed to revolve. They're not a good permanent solution for sites that serve a global audience, especially when a workable solution already exists. – Pëkka Jan 1 '14 at 18:45
  • 4
    @Pëkka The best solution is not to destroy the data, but to retain it. Percent-escaping the individual UTF-8–encoded bytes is the next best thing to trivial. That said, it is difficult to associate random squiggles like 🀄 with a particular user if you do not know what that “means”. – tchrist Jan 1 '14 at 22:16
4
+150

This is already done, but apparently not all characters are covered yet in the translation.

See Jeff's answer in Non US-ASCII characters dropped from full (profile) URL.

  • +1, I left a comment there. Hope to be fixed soon. – Ionică Bizău Jan 1 '14 at 17:06
  • 1
    Jeff no longer works at SE, so leaving a comment for him might not help, @Ionică... – Arjan Jan 1 '14 at 17:09
  • 1
    Anyway, they will see this question. Thank you for edit! – Ionică Bizău Jan 1 '14 at 17:11

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .