L2/03-280

Date/Time:    Sun Aug 24 18:52:22 EDT 2003
Contact:      peterkirk@qaya.org
Subject:      Comments on UTS #10

Proposed changes to the DUCET to change the distinction between Hebrew
final and non-final letter forms to Level 1 may be necessary for compatibility
with existing usage, although they affect only a small number of anomalous
cases. But these changes have the serious negative side-effect that, by the
normal rules, these characters will then not be, as required, treated by
default as identical for matching and searching - although they are
considered variant forms of the same letters. I have serious objections
to the proposed changes to DUCET, but will withdraw them if they are
effectively applied only to collation and not to searching and matching.

Therefore I propose that the following should be added to the list of
complications in section 8 of TR #10:

Certain characters should be treated differently for searching and matching
than for sorting. For example, Hebrew final and non-final letter forms
should be treated for sorting as distinct at Level 1, for correct handling
of some unusual cases and for compatibility with existing standards.
However, they should be treated by default as identical for searching
and matching as they are considered variant forms of the same letters.