Resource Description Format and XML-Data

Ron Daniel Jr. rdaniel at
Thu Jan 29 16:56:24 GMT 1998

At 11:31 PM 1/29/98 +1100, Rick Jelliffe wrote:

>In another forum the RDF people agreed that they *could* use the standard
>DTD syntax to markup the information they wanted. However, because they
>had a particular way they wanted the markup to look, they had to invent some
>alternative to the XML DTD syntax.

I'm sorry, but that is not correct. The RDF Schema group decided
that they would specify the schema system using the nodes and arcs
model from the RDF Model and Syntax group. This is *not* because we
wanted the markup to look a particular way. It is because:
  1) The node and arc model provides the precision for the typing
     system to build on
  2) Defining the typing system in terms of the model lets alternative
     typing systems be developed should sophisticated applications wish to
     do so.
  3) Types defined in such a way smoothly integrate with the models
     that use the types (its just more nodes and arcs).

By defining the typing system in terms of the underlying model, we
are independent of the syntax, and can come up with ways of mapping
from almost any declaration syntax (DTDs or XML-Data schemas) and
an XML instance to the nodes and arcs that are the real RDF representation
of that instance. Just how we do this is something I am
starting to look at now (as a personal effort, not an RDF group
effort). The current RDF syntax is just a first, easy, way of going
between XML and the RDF model. We want to have 'better' ways as
well, but other things have to come first.
>(Their justification for this reinventing
>the wheel seems to be "because we want it this way" rather than any technical
>reason--I am not saying they are being willful in this or that they really
>don't like the idea of markup languages; there may well be
>good reasons that they have not been able to express yet.)

I hope that the reasons above provide the technical justification
you are seeking. If the typing system were not defined in terms of the
model, we would have a real mess.

>RDF does not need a new declaration syntax. They just don't want to be

Oh, please.

>This being the case,

Its not the case, so any arguments that depend on it being the
case need other justifications.

> XML-data should not use "XML DTDs cannot support the
>needs of super-hyper-important things like RDF" as a justification for
>what they are doing.

I have not heard them use such a justification. Were they to do so,
they might lock themselves into a stronger tie with RDF than they
appear to want.

Ron Daniel Jr.              voice:+1 505 665 0597
Advanced Computing Lab        fax:+1 505 665 4939
MS B287                     email:rdaniel at
Los Alamos National Lab
Los Alamos, NM, USA, 87545  

xml-dev: A list for W3C XML Developers. To post, mailto:xml-dev at
Archived as:
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