> Mike Champion scripsit:
>
> > It would seem to me that ALL the "syntax sugar preprocessing" --
> > entity resolution and substitution, CDATA section processing, attribute
> > default value insertion, maybe XInclude resolution -- belongs
> > in some preprocessor step.
That is a matter of layering, but not of management: there needs to be
a way of naming and bundling pre- and post-parse constraints adopted.
Why should a user have to distinguish between "what elements can I use?"
and "what standard entities can I use?".
Cheers
Rick Jelliffe
