What is a namespace ... really?

[david at megginson.com]

Borden, Jonathan writes:

 > > It would cause about the same set of problems as DOCTYPE (perhaps
 > > worse with datatyping and other niceties) -- that's why we need to get
 > > away from it.
 > >
 > Ok, so you are arguing against validation in general. I like
 > how XML works now: it is optional.

Not at all -- I am a big fan of validation, and have even written a
book about XML and SGML DTDs.  

One problem with DOCTYPE is that it forces a one-to-one mapping
between the document and the DTD, and I have never worked on a medium
or large SGML system that didn't need to use different variant DTDs at
different stages of production.  To work around this problem, most
systems use shell scripts either to copy different DTD files into the
same location, or to copy different entity catalogs into the same
location.

The second problem with DOCTYPE is that the internal subset can modify 
the meaning of the external subset, so the public and system
identifers of the DOCTYPE really tell you nothing reliable about what
DTD the document conforms to.

Forcing a one-to-one mapping between namespaces and schemas would
(probably) solve the second problem, but not the first.  Architectural
forms solve both problems, but at the cost (or benefit) of adding an
additional level of indirection to the markup itself.

I've been working with SGML DTDs and SGML systems for eight years now,
and I've seen both the beauty and the shortcomings of ISO 8879.


All the best,


David

-- 
David Megginson                 david at megginson.com
           http://www.megginson.com/

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at ic.ac.uk
Archived as: http://www.lists.ic.ac.uk/hypermail/xml-dev/
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)