[
Lists Home |
Date Index |
Thread Index
]
On Thu, 2002-01-24 at 21:44, Mark Baker wrote:
> Streaming XML is a good example of where this would be a bad idea.
> If a processor hit the stream mid stride, it would have missed the
> namespace declarations. The same issue applies to processing on small
> devices, where the beginning of larger documents may be discarded from
> memory before the document is finished processing.
This kind of situation also makes it difficult to use the xmlns feature
reliable. I'm pondering some kind of mechanism for saying "these are
the namespaces I know are here, more may be lurking."
It could be because a fragment was plucked from a document, because a
document was included into another document, or because an application
simply has no idea what it assembles from various sources. In most of
these cases, a rough guideline is possible, but the rest may be
difficult.
Of course, working this way suggests the existence of "generic XML
processors", a critter whose existence many people seen to deny in
principle. (Ref: ietf-xml-mime, www-tag.) I can't figure out why.
I simply hope that the xmlns feature will give applications a
standardized means of exchanging what bits of information they do know,
based on information contained directly within the document.
--
Simon St.Laurent
Ring around the content, a pocket full of brackets
Errors, errors, all fall down!
http://simonstl.com
|
