[Foundation-l] Automatic username transliteration for SUL

Neil Harris usenet at tonal.clara.co.uk
Thu Dec 21 02:35:48 UTC 2006

James Hare wrote:
> On Wikipedia-l or something similar there has been discussion involving
> English Wikipedia's policy of blocking users with usernames that do not use
> the Latin alphabet. Reasons for opposition to this practice include
> ethnocentricism and messing up SUL. However, a point was raised in that
> people unfamiliar with the script will just see it as a bunch of
> squiggly-lines. A suggested remedy was having people transliterate their
> names depending on the wiki.
> I'm interested in combining this with a script similar to the Automatic
> Conversion script employed on the Chinese Wikipedia, that would, combined
> with SUL, automatically transliterate usernames contingent on the wiki they
> are on. For example, on French Wikipedia, your username would be in Latin
> script, whereas on the Hebrew Wikipedia your username would be in Hebrew
> script and on Arabic Wikipedia your username would be in Arabic script.
> I know there are scripts out there that can transliterate -- in addition to
> the aforementioned conversion script on the Chinese Wikipedia, there are
> quite a few scripts out there that will allow you to input something and
> have it output in a different script.
> http://vereb.free.fr/transliteration/transliterator.html has a pretty cool
> one that allows you to type with Latin alphabet settings and it will output
> in a different script. For example,
> - MessedRocker becomes МесседРоцкер with the Cyrillic setting
> - MessedRocker becomes ΜεσσεδΡοcκερ with the Greek setting
> - MessedRocker becomes مِصِدرُطكِر with the Arabic setting.
> If something could be created for SUL that would take a username and
> transliterate it depending on the language of the wiki, that would be great.
> I understand there are technical issues involved, but I would like to
> discuss it on a community level.
> --James

If we can make it as simple as that (and acknowledge that the automatic 
transliteration will often be very, very,  bad) we could possibly make 
this work, with the option to change your nick to something else later.

Perhaps language -> IPA -> language might be a good way of helping the m 
x n problem for this, in which case we could use language <-> IPA tables 
to bootstrap this.

Would this be a good idea, or would it be linguistic nonsense?

-- Neil

More information about the foundation-l mailing list