& XSLT * Petr Cimprich <petr@[...].cz>  [2005-10-18 10:15]:
>  Anyway, as we would also need a corresponding rule for
>  *_prefix_mapping methods, I would not introduce new rules
>  unless there is a strong demand and a wide consensus.
>  Completing fragmentary streams with a filter should work well
>  for now.

I had to think about this for a while, as something about it
didnâ??t sit well with me.

What I donâ??t like about it is how it forces the fragmentary
stream filter to reimplement the prefix generation stack and
algorithm for streams where the user has no particular
preferences with regard to prefixes.

If you really do not to make rules for the handlers, then the
Perl-SAX spec should at least explicitly say that behaviour in
the case of fragmentary streams is undefined; so that users know
to expect that fragmentary streams should not be expected to
work.

Otherwise you have the mess we have now: code which works with
one handler implementation fails with another, without any
indication as to whether the error is on the driversâ??s or the
handlerâ??s part.

In other words, we currently have an interoperability problem.

Regards,
-- 
Aristotle
â??Like punning, programming is a play on words.â?
   â?? Alan J. Perlis, â??Epigrams in Programmingâ?
_______________________________________________
Perl-XML mailing list
Perl-XML@[...].com
To unsubscribe: http://listserv.ActiveState.com/mailman/mysubs