With SUL launch for all users slowly coming, there's a need to clarify the global view of non-Latin alphabet usernames and resolve issues concerning their usage on various Wikimedia projects. What's the issue? Unlike the English Wikipedia, many project still have username policies that disallow names in foreign (i.e. non-Latin for most projects) scripts - so users coming from Wikipedias using non-Latin scripts might be blocked for their usernames in some projects using the Latin alphabet. This defies one of the greatest benefits of SUL - being able to use a single username over all Wikimedia projects. This issue needs to be resolved before SUL is enabled for everyone, or we can expect all kinds of complaints from disappointed users and we won't be able to do anything about it.
There are two possible solutions:
A global username policy. While this requires no technical input, numerous problems can be expected with informing local communities and their response.
A software workaround, for example, having a Latin script alias for non-Latin script usernames. This solution would probably be much easier to agree upon and use, however there's the obvious problem with technical solutions - there would have to be someone who's willing to implement the proposal into MediaWiki (and possibly maintain it later, e.g. in case projects would have to request using Latin aliases in Bugzilla), and the implementation might take a long time, possibly too long to be useful anymore.
Which one of these solutions should we focus on, or is there another one? What are your ideas and thoughts? — Timichal 22:46, 26 March 2008 (UTC)
If non Latin usernames can work on en.wp, why not elsewhere? Yann 23:10, 26 March 2008 (UTC)
There's a Wikicouncil in the works that could, given global mandate, set policy like this where the Board currently prefers not to. Otherwise, there's no global group that can set an appropriate policy. ~Kylu (u|t) 23:22, 26 March 2008 (UTC)
Imho non latin charactes should be allowed globally, so, imho, a global username policy would be a good idea. Best regards, --birdy geimfyglið(:> )=|∇ 00:25, 27 March 2008 (UTC)
I think that giving priority to one alphabet just because it is being more used is a proper idea when we define a standard such as UTF-8. But does it apply when we speak about a Wikipedia-policy? I don't speak nor read Hindu, Chinese, Arabian and many other *exotic* languages and a number of them is even shown in arrays of "?"s on my computer. Yet I don't feel uncomfortable in a communication with owner of a such username so long as s/he is doing the job well or even has provided a transcription of the username or a nickname on the user page. I start from the point that since these symbols are used to represent their own names, there should not be a reason why they should not be allowed in their global usernames. freeman-srtaLk 17:38, 28 March 2008 (UTC)
Yet, if there be no consensus about allowing people to make / use such usernames, I agree that it will be a good idea to use masks which would not change login parameters, nor be shown to the people who want to see the original usernames. So, I speak about something like "[Checkbox] Show original non-latin usernames" or "[Checkbox] Don't show original non-latin usernames, use masks/transcriptions" option. This sounds reasonable to me. freeman-srtaLk 17:46, 28 March 2008 (UTC)
What is the alterantiv to allowing people to use their non-latin usernames? With SUL I tink it will hard to stop, short of banning everyone with a non-latin usernamne who tries to use their global account at a latin wiki. And I don't think we wan't to do that :-). The best option is to have a global policy that clearly states that usernames accepted a one wiki, should be accepted at all the others. If we can get some kind of software help for those who would like translitterations of non-latin characters, I'm all for that though. --MiCkEdb 19:07, 28 March 2008 (UTC)
Implementing a global user name policy would likely be both very hard and overly time consuming. Assuming, we cannot get one:
I do not think that a global transliteration algorithm would the too hard to design for most abjads / alphabets / abugidas / syllabaries. (Note that even ideographic scripts, such as chinese, have provisions for quasi phonetic writing of foreign names and words.)
But of course, there is little or no gurantee that transliterations meet wiki requirements for acceptable names. Just think of something wonderful in one language that transliterates into a swear word in some other language.
Transliteration not an 1:1 mapping, odds are that, for some scripts, various names might get transliterated the same - specifiaclly for abjad letters such as Hebrew and Arabic ones.
I think that some users may not want their names transliterated but rather translated.
Leaving alone the question that we need a pretty universal dictionary to automatically translate names (maybe OmegaWiki will do),
there is likely little guarantee that translations do not cause name clashes from homonymes and/or homographs.
Of course, mixing transliteration with translation is going to cause yet another source of conflicting names.
Does that render us hopeless? Shall we have to overthrow any local name policies in order to get SUL working? Force every wiki into accepting any other wikies names?
No, I say, it even that is likely to cause controversies. There are quite some lookalike letters between scripts (even within one script) that - sometimes depending on the font used to display them - are either rendered identical, or very hard to distinguish. Ok, local user name policies in part forbid them, so once one name is taken, you are not allowed to register a quite different one if it is, by some software determination, a look-alike or a look-too-similar. And back we are to having to enforce user name policies globally.
Non-Latin names can cause problems for other editors, for example when trying to go to that user's talk page.--Cato 16:03, 1 June 2008 (UTC)
Latin-Alphabet "imperialism" makes me feel ill at ease, though (or because) my native French uses Latin script. Let's recapitulate a few of the above ideas, and maybe add some more.
Keep unchanged in whatever script was used. But the result might be hard for people from a different writing system.
Transliteration. Sometimes, however, there are conflicting transliterations, as with, let's say, a single set of CJK characters, but transliterated as pinyin (mainland Chinese), some other Chinese transliteration (Taiwan Chinese), romaji (Japanese) or whatever is used for Korean hanja.
Translation. Note that translation should equally well apply within the set of Latin-script languages: most "meaningful" names are not translated without a change of orthography.
UTF-8 hexcode. This one is one-to-one but not very human-readable.
Punycode. This one is sometimes easier to type, but not very human-readable either.
Relaxing the rule that all usernames of a single user must be identical, and letting a user merge existing accounts with different usernames in different wikis. But then, which username shall we "automagically assign" to user (let's say) Чижик from ruwiki when he browses one of those "Latin" wikis?