On Thu, Apr 09, 2009 at 06:59:27PM -0400, Paul Davis wrote: > Also, I still don't see anything in this damned collation algorithm > that explains how A < aa.
Maybe it boils down to something like this? - compare strings considering only primary weights - only in the case where two strings are identical sequences of primary weights, fall back to comparing secondary weights - then ditto for tertiary, quaternary Since 'a' and 'A' have the same primary weight, this would give [aA] < [aA][aA]
