W3C home > Mailing lists > Public > www-rdf-interest@w3.org > November 2001

RE: Namespaces wihtout "#" Was: Few CWM Bugs

From: Mark Birbeck <Mark.Birbeck@ePolitix.com>
Date: Wed, 28 Nov 2001 13:28:21 -0000
Message-ID: <6403181231ADB9438FE28C10F35C5B840CE2C9@S007.x-port.net>
To: "'Peter Crowther'" <peter.crowther@networkinference.com>
Cc: www-rdf-interest@w3.org
Peter Crowther wrote:
> Would current model theories deal appropriately with this if another
> fragment of RDF made an assertion about a different anonymous 
> node of class
> con:Person and with nickname "Seth"?  And what is 
> 'appropriately' here?
> Should those two nodes be merged, on the assumption that they 
> are somehow
> referring to the same thing, or kept separate?

Separate.

I've never really seen why there is a problem with the identity of an
anonymous node, since you only ever have anonymous nodes where a statement
is the object of another statement but it has no life outside of that
relationship - which means it doesn't matter what the identifier is, as long
as it plays the role of linking the two statements. And if two distinct
anonymous nodes represent the same concept and this needs reflecting in the
model, then the node shouldn't be anonymous.

But all that seems to be away from the main point, which is:

Tim Berners-Lee wrote:
> > But don't say    <> a :Person, because 
> > <http://robustai.net/~seth/index.htm>
> > is a web page and web pages are not people.

How do we know it's a web page? We only know that if we make a GET request
using HTTP and see what we get back, which seems to me to be outside of the
scope of an RDF triple. We could say we know it's a web page *outside of*
RDF because the structure takes the form of a resource *locator*, but within
RDF it is nothing more than a resource *identifier*.

I would suggest that within discussions on RDF we can only deal with URIs.
If we don't then we end up having to know more about the specifics of a
particular URL scheme. For example, the following statements would be
equivalent if we interpret the subjects as anything more than URIs:

<http://www.x-port.net/index.htm> a :HomePage
<http://www.x-port.net:80/index.htm> a :HomePage
<http://www.x-port.net/> a :HomePage
<http://www.X-PORT.net/index.htm> a :HomePage

However, I think they shouldn't be regarded as equivalent, especially since
to make them equivalent would require a plug-in to an RDF parser for every
URL scheme.

This harks back to TBL's earlier point in the discussion:

> The second issue is more significant.   In my worldview,
> (which I claim to be (a) consistent and (b) useful)
> http://example.org/x is a document.  You can't reuse
> its URI for an abstract thing without a change to HTTP.

As I've tried to argue, surely RDF can only be aware of URIs, and not URLs,
and definitely not HTTP - especially since we all agree that RDF/XML
delivered over HTTP is only one way of transporting statements, and HTTP as
a scheme for URLs, is only one way of generating URIs.

Now, whether we should agree not to use URLs when we want URIs (or more
precisely, as in this case with a :Person, a URN) - that's a different
issue. It appears to most people that I can safely generate URIs in the
namespace http://www.x-port.net/ because I own that domain. Unfortunately
that isn't true (because I can't stop someone else inventing resources in my
namespace), and as Tim says, using URLs can cause confusion. But I don't
know if we solve the problem by coming down from the abstract notion of a
URI, to the specifics of URLs.

So to sum up, I would say that using URLs as URIs does not 'break' anything
in RDF even though it may be confusing. (However, using a '#' does! Anything
after the '#' is supposed to be not part of 'the URI' as discussed in
RFC2396. So [http://www.x-port.net/a#b] and [http://www.x-port.net/a#c] are
equivalent URIs, although different URLs. Luckily, RDFMS defines the
identifier of a resource not as a URI, but as a "URI plus optional anchor
ID", or what RFC2396 calls a URI-reference. We are therefore using a
shorthand when we talk about URIs in RDF, because they are really URI
references.)

Regards,

Mark Birbeck


Mark Birbeck
Managing Director, x-port.net
Co-author Professional XML, and Professional XML Meta Data, both by Wrox
Press

E: Mark.Birbeck@x-port.net
T: +44 20 7878 1509
M: +44 7774 102 413
W: http://www.x-port.net/
Received on Wednesday, 28 November 2001 08:27:19 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:52 GMT