Lists Home |
Date Index |
- From: Paul Prescod <email@example.com>
- To: firstname.lastname@example.org
- Date: Fri, 29 May 1998 17:01:51 -0400
Tim Bray wrote:
> This distinction seems questionable. I'd like (potentially) to be able to
> issue a query along the lines of "find me elements in this repository
> which have a creation-date prior to 1998 and which are of a type that
> can have ProjectNumber as a child element."
Here is how I understand Tim's argument:
Just as the benefit of using XML as an encoding for XSchema is that code
and even full applications that have never heard of XSchemas can do
interesting things with XSchema documents, the benefit of using RDF as an
encoding for XSchemas would be to allow RDF-smart code and even
applications to do interesting things with XSchema documents without
specifically knowing about XSchema.
If this is the important argument, then the "correct" level of RDF use is
the level at which this sort of "blind underlayer use" is practical. For
instance it makes perfect sense to ask about which element types have a
HasAttribute property with value "xml:link". It also makes perfect sense
to ask about which documents have a ConformsTo property of "HTML 5.0".
But when you get into trying to do queries on content models, you quickly
run into a problem. It's easy to check the
[element] ---CanContain--> [ProjectNumber]
relationship. That's essentially just asking if ProjectNumber is anywhere
in element's content model.
But try checking that an element type can contain BOTH a ProjectNumber AND
a ProjectDescription. That's easy with regular expression theory, but I
don't think it can be expressed in RDF alone. The relationship is more
complicated than a property/value relationship.
This suggests to me that RDF breaks down at that level. I think that RDF
should be used to:
* describe the relationship between documents and schemata.
* describe the relationship between elements and their attributes.
But I can't see an intuitive RDF-compatible encoding for content model
relationships. Some types of relationships are just outside of RDF's
RDF-philes, do you agree?
Paul Prescod - http://itrc.uwaterloo.ca/~papresco
Three things see no end: A loop with exit code done wrong
A semaphore untested, and the change that comes along
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)