Ok. It's all hanging under an object tree with a classID in the
toys they give me at work. Sigh.. we are MSThralls and while it
may not be kosher, the code runs. :-)
If wrong, kick me. An ns attribute addresses a schema. That
schema defines the namespace. In some examples, that points to a
DTD, doing the job a Doctype does without the other stuff and
no assertion about the capability of the schema except that
it can be used to disambiguate the namespace for properties
in the local scope. In others, it points to a classID in a registry.
One might be semi-sorta-kinda-liftTongue BNF, and the other is, here in
ThrallLand, an Active-X object.
So it looks to me like both models are right and both
represent different things. For the purposes of creating
a compound instance, namespaces look good. Validation?
Having one file with multiple parse/validation contexts
seems to go way beyond the skill of the DePH, and is
maybe HarderThanSGML????? Where are those simple requirements?
Ok, what IS a compound document type definition (CDTD)?
It is a very interesting problem. Here is my challenge:
regardless of public relations and so on, an SGML DTD is a
reasonable thing to look and use to eyeBall parse a
file given resolution of the parameter entities. It
takes a little practice, but most oldSGMLers can do it.
An XML compound DTD should also provide this ability.
len
PS: I promise to read XML-Data ASAP.