Again wit da AND and Repetitions

roddey at roddey at
Mon May 17 20:40:24 BST 1999

>I don't remember SQL ever adopting the view that the only integrity
>constraints you were allowed to specify were those that could be evaluated
>in linear time. In fact, the refusal to build implementation-based
>limitations into the language was one of the major reasons for the success
>of SQL.
>(In implementing GedML I discovered that the integrity constraints that I
>could specify in the DTD were such a pathetic subset of the total that I
>might as well do all the validation in the application and ignore the DTD
>capabilities entirely - especially as I had no way via the SAX API of
>knowing whether the parser had done any validation or not).

I guess my argument though is that whatever you put in, its either too much or
too little. Therefore, you should have a small, fast and compact set of built in
constraints and provide for an escape mechanism for more complex constraints. I
guess the argument against your argument would be that SQL is still an extremely
limited mechanism. XML will be much more widely applicable and even if you made
it ten times as large, complex, and piggy, you still won't handle a major
fraction of the constraint requirements (in terms of structure and data typing)
that the whole world will want, so some sort of layered escape mechanism is
desirable no matter what you do. So why put in stuff that will be very heavy and
for which everyone will pay when it still won't be enough? Keep the core fast
and simple and applicable to what it was originally designed for, and push the
heavy stuff out to the periphery or the application. I don't consider Schema to
be the periphery, I consider it to be something that will become a core piece of
the system, so everyone will pay for any pigginess in it.

Just my opinion of course...

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as: and on CD-ROM/ISBN 981-02-3594-1
To (un)subscribe, mailto:majordomo at the following message;
(un)subscribe xml-dev
To subscribe to the digests, mailto:majordomo at the following message;
subscribe xml-dev-digest
List coordinator, Henry Rzepa (mailto:rzepa at

More information about the Xml-dev mailing list