[
Lists Home |
Date Index |
Thread Index
]
Hi Elliotte,
> Fortunately DOM 3 does not require using the processing model
> implied by W3C XML Schema and DTDs. If you've got another processing
> model, go for it. All you need to do is map the types you do assign
> to namespace URIs and type names. I'm not intimately familiar with
> RELAX NG, but I suspect a RELAX aware processor could easily
> generate a typed DOM. It's just that two different RELAX processors
> might not build identical DOMs when it came to type information.
>
> The only problem would arise if RELAX wanted to assign multiple
> types to the same element or attribute when building a single DOM,
> or if its types somehow did not match the element and attribute
> structure of the XML document (though I'm not sure that's a good
> idea in any case.)
I think that DOM 3 can support these annotations as well, using the
get/setUserData() methods on Node. If RELAX NG, Schematron or other
processors want to associate other information with a node, such as
more than one type or diagnostic messages, then they can do so through
that interface. All that would be needed would be to standardise what
keys to use for particular types of information. (This is the kind of
flexible annotation that myself, Uche and others have been discussing
for XPath NG.)
Cheers,
Jeni
---
Jeni Tennison
http://www.jenitennison.com/
|
