W3C home > Mailing lists > Public > www-rdf-interest@w3.org > November 2000

Re: Using XHTML as an RDF Language

From: Sean B. Palmer <sean@mysterylights.com>
Date: Sat, 25 Nov 2000 18:57:50 -0000
Message-ID: <008801c05711$9d901560$cf4e01d5@z5n9x1>
To: <www-rdf-interest@w3.org>
Cc: "Dan Connolly" <connolly@w3.org>
{Mr. Connolly: You may be interested in [2]}
{Everyone else: This is further to my original note [1]...}

You may point out that we may still need to use the class mechanism for
triplets etc. that can't be described in the XHTML namespace:-
     <span class="subject">This page</span> was
     <span class="predicate">written</span> by
     <span class="object">Sean B. Palmer</span>
But then triplets are subjects of the pure RDF syntax model, whereas the
<address> tag is a part of the XHTML namespace and needs further semantic
description. Bascialy, you could state that address may contain triplet
assertions about a pages author, or something similar, where you could
define that the subject and predicate were automatically "this page" and
"author" respectivly, so that the only further part of the triple you would
need is the object:
     Sean B. Palmer - mailto:sean@mysterylights.com

See my Notation3 hacking for this [2]

Anyway, RDF processors looking up the contents of the <address> tag would
look at the RDF Schema (or Notation3!) for XHTML and automatically realise
that the contents are an object property for the author of the given
The same kind of thing applies for <title> where the contents are the object
for the triple:
[ subject = this page | predicate = title | object = (contents of <title>) ]
And by using this logic you could apply to similar elements.
What about <p> etc.? Well, you could just write a small prose description
stating that there is no given RDF property for it, except that it acts as a
paragraph language container, and may contain whatever <p> may contain
structurally. Further semantics in that would have to be added with <span
class="[...]"> or direct RDF (at a later date).

[1] http://lists.w3.org/Archives/Public/www-rdf-interest/2000Nov/0381.html
[2] Hmmmm.....in Notation3 [3] this might be expressed as (probably
incorrect: you'd have to embed it in a page somehow):

bind dc:  <http://purl.org/dc/elements/1.1/>
bind html: <http://www.w3.org/1999/xhtml>
bind default  <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
  [ object "html:address" ] is dc:author of [ a subject; about "this page" ]

- @@ Can "about" be left blank to signify "this page" in RDF, or do we use
the GenID for that? Could we use XPointer for html:address? I haven't really
looked in to Notation3, so what I have written is probably just a pile of
garbage. Note that using Dan Connolly's Python N3=>RDF converter, the
results are:-
<web:RDF xmlns:web="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
<web:Description about="http://pageexample.org/1">
 <web:Description about="http://pageexample.org/2">
  <type xmlns="http://www.w3.org/1999/02/22-rdf-syntax-ns#"
  resource="http://www.w3.org/1999/02/22-rdf-syntax-ns#subject" />
  <about xmlns="http://www.w3.org/1999/02/22-rdf-syntax-ns#">this
  <author xmlns="http://purl.org/dc/elements/1.1/#"
resource="http://pageexample.org/1" />
If that were embedded into a page, "this page" would make sense I suppose

[3] http://www.w3.org/DesignIssues/Notation3

Kindest Regards,
Sean B. Palmer
"Perhaps, but let's not get bogged down in semantics."
   - Homer J. Simpson, BABF07.

----- Original Message -----
From: Sean B. Palmer <sean@mysterylights.com>
To: <www-rdf-interest@w3.org>
Cc: <xml-dev@lists.xml.org>; <swi-dev@egroups.com>
Sent: Saturday, November 25, 2000 4:56 PM
Subject: Using XHTML as an RDF Language

> If we want to provide an RDF view of an XHTML document, at the moment we
> the class attribute in XHTML and convert into DC and RSS. However, if
> is a semantic XML description language that contains all of the RDF
> information in the first place, couldn't someone write an RDF Schema for
> XHTML? I already wrote one for the profile attribute [1], but wonder if it
> can't be expanded to cover all of XHTML?
> We may have to wait for an XML Schema version of m12n, to sort out some of
> the cruft...but otherwise it's not a bad idea: using XHTML as an RDF
> language in the first place, rather than scraping all of our RDF data out
> it [2].
> [1] http://lists.w3.org/Archives/Public/w3c-wai-er-ig/2000Nov/0057.html
> (came out a bit odd)
> [2] http://lists.w3.org/Archives/Public/www-rdf-interest/2000Mar/0103.html
> BTW: My email is crashing completely at the moment, so if you send
> replies, I might not get them for a while.
> Kindest Regards,
> Sean B. Palmer
> http://xhtml.waptechinfo.com/swr/
> http://www.w3.org/WAI/ER/
> http://www.w3.org/WAI/GL/
> "Perhaps, but let's not get bogged down in semantics."
>    - Homer J. Simpson, BABF07.

More hacking. At least this one kind of works...
bind dc:  <http://purl.org/dc/elements/1.1/>
bind rdf:  <http://www.w3.org/1999/02/22-rdf-syntax-ns#>
  bind default <http://xmlns.com/foaf/0.1/>
  [ givenname "Sean"; surname "Palmer" ] is dc:author of
  [ a rdf:subject; rdf:about "http://www.waptechinfo.com" ]
Received on Saturday, 25 November 2000 13:58:34 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:07:33 UTC