> > It's actually up to four array plus some extra maps.

I have found ICU (International Components for Unicode)
site and it contains a lot of information I should have
been aware of before plunging in here.

Here is the link:

http://oss.software.ibm.com/icu/

and this is the bit that seems to be relecant for this
thread (single line)

http://oss.software.ibm.com/icu/userguide/Collate_Customization.html

BTW, the code is said to be available in Java, C, C++
under a non-viral version of GPL.

Cheers,
Adem