W3C home > Mailing lists > Public > www-rdf-interest@w3.org > January 2004

Re: [www-rdf-interest] <none>

From: Charles McCathieNevile <charles@w3.org>
Date: Tue, 13 Jan 2004 08:04:20 -0500 (EST)
To: Jeroen Budts <jeroen@lightyear.be>
Cc: www-rdf-interest@w3.org
Message-ID: <Pine.LNX.4.55.0401130756370.17545@homer.w3.org>

On Mon, 12 Jan 2004, Jeroen Budts wrote:
>I'm still working on my FOAF file and have another question. To give
>some extra information about the persons I know (foaf:knows) I want to
>add XFN [1]. XFN describes some values which can be used in the @rel
>attribute of XHTML elements. I know the rel attribute, as XFN uses it,
>is http://gmpg.org/xfn/1#rel. So is it valid then to declare a prefix,
>by example xfn, for the namespace http://gmpg.org/xfn/1 and use
>xfn:rel as a property for a foaf:Person?

Yes, I think you can do this. It seems to be what the document you referred
to suggests in its last paragraph. (Assuming that it is a sensible use of the
rel property :-)

An interesting problem is whether to adapt something like this, or to
re-define it yourself. If you take the former appraoch then you either don't
write a schema, so people can't find what you did, or you have the
possibility that several people will each write a schema for a term whose URI
they don't own, and there may be conflicting statements.

If you make up your own version, we end up with a million schemas which are
likely to have ppor interoperability - even with the core elements of dublin
core we can see that establishing interoperability among a large group of
humans is difficult. If they all invent their own somewhat nuanced terms and
claim they are "similarTo" something else, we will have a very fuzzy semantic
web. That may be a good thing of course - the knowledge we try to represent
is, except in a few cases, somewhat fuzzy itself.


Received on Tuesday, 13 January 2004 08:04:20 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:07:49 UTC