Lists Home |
Date Index |
- From: David Megginson <firstname.lastname@example.org>
- To: "email@example.com" <firstname.lastname@example.org>
- Date: Wed, 1 Apr 1998 07:12:47 -0500
james anderson writes:
> 1. when a namespace-pi binds a namespace, is it intended that,
> should a schema have been specified, a processor verify
> (immediately?, later?, when?) the existence (the content?) of the
> specified schema? is this a well-formedness or a validity issue?
This is currently undefined -- think of the current namespace WD as a
hook for future enhancements rather than a self-contained spec.
Namespaces can be neither a well-formedness nor a validity issue with
XML 1.0, though the WG might decide to change that in a future version
Personally, I'd prefer to keep namespace processing as a completely
separate layer (like TCP on top of IP or HTTP on top of TCP), so that
nothing in namespaces will ever affect the basic validity or
well-formedness of XML documents. Imagine if every change to HTTP
required a change to IP as well!
> 2. if the schema is present, should the processor permit local
> additions to the namespace, that is the introduction of names which
> are not present in the external definition? should the processor
> permit redefinition of existing names from the namespace?
This would go against the basic principle of namespaces (globalisation
and uniquification of names), since two documents could create
different extensions to the same namespace. Of course, there's no
standard mechanism to verify this right now.
I'm not certain that I understand the issue here -- why would someone
not bring additional element types in from a different namespace,
instead of adding private extensions to an existing one?
> (or rather, it's almost possible: there's a small problem, that the
> wd-standard precludes qualified entity names. why?)
The namespace spec allows element type names, attribute names, and PI
targets to be associated with a URI. (External) entity and notation
names are already associated with a URI.
> 3. the element definition examples below shouldn't, in any event,
> appear in the original schema(s). while it is ok (and necessary)
> to constrain the namespace in definition tags in the internal
> subset, to do so in the original schema itself would prevent
> subsequent users of the dtd from remapping the tags to suit their
> needs. in general this is too restrictive.
Absolutely correct. Currently, XML documents have only one schema
language -- DTDs -- and each XML document may have only one of those.
Nothing in the namespaces WD (or anything else) is allowed to override
XML 1.0, which is already a W3C recommendation. However, if DTDs are
used as namespace schemas in the future, then I would assume that the
prefixes will _not_ be hardcoded in those schemas.
All the best,
David Megginson email@example.com
Microstar Software Ltd. firstname.lastname@example.org
xml-dev: A list for W3C XML Developers. To post, mailto:email@example.com
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:firstname.lastname@example.org the following message;
To subscribe to the digests, mailto:email@example.com the following message;
List coordinator, Henry Rzepa (mailto:firstname.lastname@example.org)