Localization in code

Jecel Assumpcao Jr jecel at merlintec.com
Fri Sep 28 14:54:03 UTC 2001


On Thursday 27 September 2001 21:15, Richard A. O'Keefe wrote:
> Ideally, we'd have a natural language generation system rather like
> ILEX so that no translation as such would be involved at all.  That
> kind of stuff is getting _very_ close to ready for use.

Any reference for ILEX?

I've posted my comments about localization problems before (sadly, the 
Yahoo archive doesn't go back that far. Here is a reference that loads 
1MB - http://www.create.ucsb.edu/squeak/9710.html#Letter406). I was 
more worried about translating the programming language itself rather 
than just strings, but my solution would handle both.

Since then, the Universal Network Language project has done most of the 
hard work needed:

http://www.ias.unu.edu/research_prog/science_technology/universalnetwork_language.html

It is a pity that such a UN funded project isn't more open.

If such a system could be included in Squeak, then we could have 
special UNLString objects used instead of regular Strings and their 
#printString method would generate ("deconvert") the message in the 
current natural language. But if we were to inspect it using any system 
tool, it would still have all the UNLString information and could be 
edited in special ways.

-- Jecel




More information about the Squeak-dev mailing list