On Tue, Aug 17, 2010 at 4:45 PM, Nikola Smolenski smolensk@eunet.rs wrote:
All right; but right now we are not paying attention to character context too, and not properly sorting even single characters. I mean, we are sorting Ђ before А! Surely this would be an improvement?
Yes, but it would probably be easier to just use CLDR or ICU to begin with, rather than trying to write and debug our own sort function.