W3C home > Mailing lists > Public > www-rdf-logic@w3.org > August 2001

Re: Summary of the QName to URI Mapping Problem

From: Jonathan Borden <jborden@mediaone.net>
Date: Sun, 12 Aug 2001 13:08:07 -0400
Message-ID: <002901c12352$14ada7e0$0301a8c0@ne.mediaone.net>
To: "Piotr Kaminski" <pkaminsk@home.com>, "pat hayes" <phayes@ai.uwf.edu>
Cc: <www-rdf-logic@w3.org>
Pat,

For what it is worth, I think this is a major inconsistency between XML and
RDF. As the purveyor of both specifications, the W3C really ought ensure
that these two areas be consistent, but there have been areas where they are
rather inconsistent on some of these somewhat important issues. Case in
point, it is not -currently- possible to -name- a general XML Schema type
with a URI, and furthermore XML Schema does not see URIs as the naming
mechanism for types, rather QNames.

So you see that even in rather recent W3C specifications, URIs are not used
as names. I suggest that this lends evidence to my position that the XML and
RDF communities have different ideas about the usages of QNames and URIs.

1) The XML community in general finds it unacceptable to end XML Namespaces
in '#' -- perhaps one might discover a -single- well known XML
Namespace -outside- of RDF applications that fit this pattern.

2) simply concatenation of QName namespace URI + localname is frought with
many many problems, and IMHO does serious damage to RDF's ability to
interact with XML Namespace aware applications. For example if XML Query
applications are XML Schema aware and allow comparison of xsd:decimal
values, values types with QNames whose namespace ends in '#' will not be
correctly recognized as XML Schema datatypes.

At Extreme Markup 2001, Jonathan Robie gave a -terrific- demonstration of
using XML Query to query RDF, to the point where this may now become my
preferred RDF query language!


>
> This seems to me to be the basic issue: who has the authority to say
> they are distinct?
>

Exactly. The point is that large numbers of XML people have problems with
this. Particularly XML Schema does not equate two QNames that happen to
yield the same RDF URI mapping (concat) as being equivalent.

What happens is that if RDF uses these rules it majorly looses inferencing
power because it looses the ability to understand XML Schema type
hierarchies etc. etc.

> Suppose someone were to reply: you can SAY that, but if you do say
> it, you are uttering a contradiction, since I can prove you wrong (by
> mapping the Qnames to URIs using the published RDF rules). After all,
> you can *say* (P and (not P)) as well, but I'm not obliged to believe
> you if you do.
>
> >foo:barcat
> >bar:cat
> >
> >Yet, using the standard concatenation mapping, both QNames are mapped to
the
> >same URI:
> >
> >http://example.com/foobarcat
> >
> >While this URI identifies some resource (by definition), it cannot
identify
> >both of the (distinct!) resources identified by the two QNames
> >simultaneously.  Hence the mapping is deficient.
>
> If they really were distinct, that is. Or, you could take the
> position that the use of the mapping shows that they couldn't have
> been distinct.

Well, XML 1.0 says they are distinct types. So you loose this basic
information. Your choice. I, for one, will be working with XML information,
and am looking for a system that will be able to perform inferencing,
queries etc. on such information.

-Jonathan
Received on Thursday, 16 August 2001 22:16:03 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:52:40 GMT