> I'd argue that any spec that imposes additional constraints on
> processors not imposed by XML 1.0 is not compliant with XML.
I think this might usefully be called "strict compliance". That
should be a familiar term for many people, with that meaning.
As in: "Such a spec isn't strictly compliant with XML, although
it does use XML in its surface syntax."
The general issue is what someone pointed out: what to do
with things that aren't explicitly mentioned in a specification.
A strict interpretation of any spec says that you can't ever rely
on such things ... ergo, it's as if they're forbidden. It's always
safe to read a good spec strictly. If such a reading gives
nonsense, it's a spec problem.
- Dave
