[
Lists Home |
Date Index |
Thread Index
]
Which is a shame. Having, for example, to rewrite the XHTML schema just
so I can include SVG fragments makes building compound documents almost
impossibly difficult. I suspect we need to think about more types of
validity than those currently offered by DTDs, XML Schemas, Schematron,
and RELAX NG.
Two types of partial validation that would be useful for constructing
compound documents are:
1) Validate the language you know and ignore fragments headed by unknown
elements.
2) Start at the top and only start validating when you hit an element
you know. (The opposite of #1.)
I suspect that, with a bit of thought, people can find other use cases
and types of validation to fit them that are not met by current definitions.
-- Ron
Bob Foster wrote:
> Roger L. Costello wrote:
> [snip]
> > I interpret Murata's slides as saying that with NVDL you can insert
> elements
> > from another namespace into the <Book>...</Book> element.
>
> I don't. Sticking arbitrary namespaces into an element would fail to
> validate against the schema unless it specifically allowed for such
> insertions, e.g., through wildcards.
>
> What NVDL should do is allow you to validate inserted-namespace
> subcomponents that _are_ otherwise valid but are not fully described by
> the schema for Book.
|
