Chris Fynn responded:
> Keld Jørn Simonsen wrote:
> > As you would originally only do the folding to lowercase
> > as preparation for a comparison, it is only natural to skip
> > this lowercasing step completely and just do the comparison.
> > For this to work you need an API that can do case-insensitive
> > comparison, based on 14651 data (as Ken indicated).
> > Keld
> Unfortunately ISO 14651 does not yet cover all the characters
> in Unicode 3.0 or ISO 10646 2000 part 1.
This is true, but to be fair to 14651, the number of characters
that have casing behavior added between Unicode 2.1 and Unicode 3.0
is a mere handful. Tailoring a 14651 table to add these few characters
is not a large task. But it does have the drawback of having to depend
on a tailoring rather than the unaltered table. (The same applies to
the Unicode Collation Algorithm tables.)
The recommendations for case folding in UTR #21 Case Mappings are
designed for the Unicode 3.0 repertoire, so do not have the
disadvantage of incompleteness.
This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:21:00 EDT