11/1/2002 5:39:41 PM, Richard Tobin <> wrote:
>
>I think it was right to separate entities - a lexical mechanism - from
>schemas. Character substitution should happen at a level below schemas,
>and should be usable with any schema language (and without any).
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. Then parsers per se can be clean lightweight things
that don't understand all this stuff (and machine-produced XML can write
for), whereas human-authored XML can go through one or more of these
preprocessors to get it into shape for parsing.
