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

Re: RDF in XHTML

From: Sean B. Palmer <sean@mysterylights.com>
Date: Mon, 16 Apr 2001 18:37:01 +0100
Message-ID: <02c901c0c69b$e02dc3a0$e2d893c3@z5n9x1>
To: "Seth Russell" <seth@robustai.net>, "Joshua Allen" <joshuaa@microsoft.com>
Cc: "Murray Altheim" <altheim@eng.sun.com>, "Danny Ayers" <danny@panlanka.net>, "RDF Interest" <www-rdf-interest@w3.org>
[Terse answers first, long answers second]

> So I guess the problem is, How do we factor the XHTML
> validations such that the semantics of RDF become opaque?

Yes (that's the problem).

> ... or am I missing something again ...

No (you never have been).

> But now, I'm beginning to see the wisdom of just putting a
> link in the document to an external RDF description file.
> That would solve this dilemma,  wouldn't it?

Yes (theoretically).

> Could you refresh our memory of the status of that proposal?

No (<grin/>).

But I can refresh your memories of the nature of the proposals.
Basically, there are two legal methods of linking to something in
XHTML: using <a href=""> or <link href=""/> (I guess one could use
"<object>" as well, but what's the point?). <a> is good because people
can just click on it, but it is hardly machine retrievable. <link/> is
better because machines often already scrape these links out (e.g. for
stylesheet information), but the drawback is that it uses "link types"
to identify the type of content being linked to, and there is no
widely agreed upon default (i.e. given by the HTML specification) for
identifying metadata. There are various hacks around this, but what
people require is some heavily agreed upon and stably implementable
system. This ain't gonna happen, unless you use something like RDDL.
In other words, where I answered "yes" to linking to RDF data solving
our problems, I meant in an ideal world, not the current practical
one.

Thus, I guess I conclude that embedding information into XHTML is
quite a good idea, when you weigh up all sides of the debate. The
problem once again is that the RDF data model is very dissimilar from
the XHTML data model, because one is metadata oriented, and one is
data oriented. Adding one to the other at this stage is something that
is going to have to be controlled to avoid numerous proprietary
implementations: i.e. adding RDF to XHTML should be well scoped rather
than "put your RDF here".

I suppose that the aim is "put your RDF application here" - i.e. if
you're embedding your RDF into XHTML, you're going to be doing it for
a *reason*, and more often than not, that reason is so that something
can process it. I don't know of anyone who wants to put RDF into XHTML
just for decoration :-)

--
Kindest Regards,
Sean B. Palmer
@prefix : <http://webns.net/roughterms/> .
:Sean :hasHomepage <http://purl.org/net/sbp/> .
Received on Monday, 16 April 2001 13:35:25 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:51:48 GMT