...not to forget that the IBM Classes for Unicode (open source) in their version
1.3 now include an implementation of the bidi algorithm, too.
this one is probably slower than mark's "pretty good bidi algorithm", because it
conforms to the version of the algorithm that is part of unicode 3.0 (that is
version 5 of the tech report).
if you need conformance, then you may have a look at it.
the author ( :-) spent some time tuning the performance, and it tests 100%
against the reference implementation that is going onto the unicode 3.0 cd.
ibm classes for unicode:
Markus Scherer IBM Cupertino, CA email@example.com
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:53 EDT