XML complexity, namespaces (was WG)
mrc at allette.com.au
Sun Mar 21 22:34:40 GMT 1999
David Megginson wrote:
> In the SGML world, someone had to create one or more variant DTDs to
> ensure that the document was valid at each stage (either through
> elaborate and obfuscatory configuration with parameter entities and
> marked sections, or by writing separate DTDs from scratch, both often
> with liberal sprinklings of ANY). If the production engineers made
> even minor changes in the system design, the DTDs would (usually)
> immediately break validation and shut down the whole system -- in
> other words, the system was brittle and expensive to maintain.
> Sometimes, the benefit of tight validation at each step makes the
> extra expense worthwhile, but as I mentioned before, XML lets you make
> that decision rather than making it for you.
I think it becomes difficult to continue discussing this issue as we move further into
theoritical datasets and workflows, as we are probably all forming a mental picture of a
scenario that serves our own positions. While I agree with much of the above, I think it's
only fair to point out that the variant DTDs need not be created just for the sake of
validating the instances in their current state. They also provide context for the next
conversion, which can reduce the coding required to get to the next stage. In that role, they
can save money by simplifying the coding and accurately documenting the structure of the
instances. If the changes are simple, then so might be the maintenance of the DTDs - if
they're complex (such as converting specialised tables to CALS tables), the task can be made
much easier if you have access to the structure. It could be argued that changing the order of
the six transformations that you provided would be easier if you had a DTD describing the
structure at each point, making it more expensive up front, but potentially more flexible; at
least you might not have to start from scratch.
Of course, all of these scenarios depend on the dynamics of the project - I just don't think
that there is an automatic and substantial loss associated with validation at all stages. I do
agree that it's good that XML allows you to make the choice about validation, but I don't
accept that it will always be cheaper.
Marcus Carr email: mrc at allette.com.au
Allette Systems (Australia) www: http://www.allette.com.au
"Everything should be made as simple as possible, but not simpler."
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/ and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo at ic.ac.uk the following message;
To subscribe to the digests, mailto:majordomo at ic.ac.uk the following message;
List coordinator, Henry Rzepa (mailto:rzepa at ic.ac.uk)
More information about the Xml-dev