Lists Home |
Date Index |
> From: Marcus Carr [mailto:firstname.lastname@example.org]
> Sure, but I doubt if any v1.1 parsers would bother to support
> DTDs, given that DTDs
> would suddenly only be capable of expressing a subset of the
This is totally bogus. Current market-leading XML parsers have support for
XML Schema and namespaces, as well as supporting DTDs. Your argument that
vendors will abandone DTD support if XML 1.1 rolls in namespace support
flies in the face of what vendors have already done. There is nothing to
support such an assertion. It is simply unfounded paranoia.
> Additionally, I'm not keen to see a situation where the
> version of my document is
> dictated by whether I want to parse with a DTD or a schema.
> This just seems like the
> latest in a string of moves to get rid of DTDs, despite the
> fact that they fit very
> well in many current systems.
I can use DTDs or Schemas with the same set of tools. I do so routinely. I
use documents with namespaces and documents without namespaces, and do so
with the same set of tools without any changes in configurations. The only
constraint is that any use of ":" in names must comply with XML Namespaces.
I really find it hard to believe that this is such an onerous constraint
that we must preserve a problematic schism in current XML standards in order
to avoid imposing such a constraint on name productions in XML 1.1.
> > I'm talking about mandating their inclusion in the next version of
> > the core XML recommendation, not requiring their usage in XML
> > version N.
> The two seem inextricably linked, aren't they?
No, they are not. Not only do I routinely use documents with and without
namespaces (as mentioned above), but I have even used documents that include
namespace-scoped elements and elements in no namespace in the same document.
I had no difficulty with this. Where is the problem?
> Namespaces do not always exist, so you can't use any random
> XML parser for any random
I use any random standards-compliant XML parser for any random use routinely
-- and I've never had problems. Of course, the moment I encounter a document
that does not comply with XML Namespaces, but does use ":" in names, this
will no longer be the case. But that problem has not arisen for me, yet.
As far as I can tell, this really all boils down to simply mandating that
any use of ":" in names in XML 1.1 must comply with XML Namespaces. There is
no other cost to existing documents. I'd be genuinely interested in hearing
of real-world use cases that would warrant preserving the ability to use ":"
in names in a manner that does not comply with XML Namespaces.