Re: Text and Anti-Text

From: peter_constable@sil.org
Date: Fri Aug 13 1999 - 12:32:04 EDT


       Dean Snyder wrote:
>I propose that 10646 reserve the 32nd bit as a flag bit
       signifying text versus meta-text...

       Welcome to the list!

       I think this is an interesting proposal. I've thought there
       would be definite advantages to being able to clearly
       disambiguate text and meta-text directly in the encoding.

       I see a serious problem in your proposal, though: it works only
       for UCS-4 (assuming a redefinition to support bit 31 = 1). Most
       people are likely to work with text stored as some variant of
       UTF-16 or as UTF-8. UTF-16 has no way to represent characters
       above x7FFFFFFF; neither does UTF-8, at least as supported by
       Unicode (I haven't seen ISO's specs on UTF-8). So, while your
       suggestion may have some merit in terms of the benefits it
       would achieve, it has a serious limitation in that there is no
       way to incorporate it into the way most people will actually
       implement the standard.

       I don't want to keep you from looking for solutions to achieve
       the benefits you desire. (And I dearly hope this negative
       feedback on your first foray onto this list doesn't turn you
       off from further interaction.) I'm just not sure this solution
       would work.

       Peter



This archive was generated by hypermail 2.1.2 : Tue Jul 10 2001 - 17:20:51 EDT