[
Lists Home |
Date Index |
Thread Index
]
Richard Tobin wrote:
>Making undefined entities a validity error in order to allow
>experimentation would reduce interoperability. I don't really see any
>advantage to it over just doing those experiments with non-conforming
>processors. You don't need the XML Core Group's permission to do that :-)
>
>
I have been "experimenting" with non-conforming processors for several
years:
our products include a backwards parser, reference-ignoring parser
modes, parsers
that recover from unbalanced tags, and parsers that recover from non-WF XML.
It seems we are at an impasse here: people will not move to non-standard XML
and this is hindering them from fully utilizing tools firther down the
line (XSLT,
schema languages, XInclude, etc). But the XML Core WG probably won't
make any changes unless there is a grassroots movement of people who
have adopted some non-standard change.
So I think that Richard is sadly correct here: people who want to use
named character references and get rid of DTDs must implement their
own non-standard systems. So we should not waste our time with
standards conformance, but just go ahead and do it; if enough of
us implement the same thing, maybe W3C will add it. If a
standard is an "agreement" that reflects industry practise, I guess
that is the way it has to be. What should XML-DEVers do? Define some
SAX property to handle this, then get it implemented in the standard
SAX parsers?
Cheers
Rick Jelliffe
|
