W3C home > Mailing lists > Public > www-tag@w3.org > February 2002

Re: [namespaceDocument-8] 14 Theses, take 2

From: Tim Berners-Lee <timbl@w3.org>
Date: Wed, 27 Feb 2002 12:02:17 -0500
Message-ID: <03ab01c1bfb0$830ef770$0301a8c0@w3.org>
To: "Paul Prescod" <paul@prescod.net>, "TAG" <www-tag@w3.org>
IMHO

----- Original Message ----- 
From: "Paul Prescod" <paul@prescod.net>
To: "TAG" <www-tag@w3.org>
Sent: Monday, February 18, 2002 6:58 PM
Subject: Re: [namespaceDocument-8] 14 Theses, take 2


> Dan Connolly wrote:
> > 
> >...
> > Thesis 13 (and 14) seems to say that it's necessary to pick
> > one data format for all namespace documents. Not so;
> > to each his own. And if my namespace is pretty special-purpose,
> > what's wrong with using a special-purpose data format
> > to document it?
> 
> There is the theory floating around that namespace documents should be
> machine processable in the sense that code could go there and get
> something useful to figure out how to process a multi-namespace
> document. That would require a predictable way to get machine readable
> information about of namespace documents.
> 
> I personally feel that we should not try to put machine-readable
> information there until there is a model for how the machine would
> actually go about looking information and using it for a particular
> domain. 

The RDDL spec contains an algorithm for looking for information,
and so allows automatic validation fo a document.

The RDF and RDFS specs togther provide information as to how
to pick up RDFS information (directly or indirectly)
and use it to validate a document, or for that matter to generate
a human-readable version of the document using labels from
the schema.

I have running code which uses the RDFS algorithms to
check consistency of a document with its RDF schema, doing
recursively checking the validity of any schemas encountered too.
I use it.  It wouldn't work if I have to do some human reading of the
namespace document at each point.  ((It doesn't grok RDDL either,
and I'm not really that keen to add a rddl-to-rdf client module on
because for me it means code bloat, when rdf would have not.))

So  "a predictable way to get machine readable
information about of namespace documents" exists, doesn't it?

>For instance it certainly won't work to use the schema for SOAP
> and the schema for XHTML to generate a meta-schema for SOAP+XHTML.
> Putting a couple of RDF links in an XHTML document doesn't fix the
> fundamental problem that people combine namespaces in unconstrained and
> contradictory ways.

The problem of defining how namespaces can be mixed is a separate one
and interesting though it is I don't want to  get into it under this topic.
The problem does not exist as such with RDF, because you can use
as many RDF vocuabularies in a document as you like.  The schemas
have ways of putting contsraints on what makes sense. 
They aren't syntactic constraints.

Tim

>  Paul Prescod
> 
Received on Wednesday, 27 February 2002 11:57:08 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Thursday, 26 April 2012 12:47:04 GMT