W3C | Semantic Web Development

Site Summaries in XHTML

We use a constrained dialect of XHTML to express channels in the RSS sense:

This dialect also includes the popular convention for linking to additional RDF metadata such as Dublin Core Subject, Description, Date, Format, Language, Creator, Publisher, and Rights:

<link rel="meta" href="Overview-about.rdf" />

This dialect (more formally, metadata profile as specified in section 6.12 Link types of the HTML 4.0 specification) should be declared explicitly ala <head profile="http://www.w3.org/2000/08/w3c-synd/#">.

The Transformation in XSLT

Share and enjoy:

The W3C at meerkat channel seems to be working (Sep 2000). Keep your fingers crossed!

Variations on RSS

In an RSS-DEV discussion, Teun Duynstee contributed edits to support RSS 0.92. Danbri integrated them (2001-05-31) into home2rss092.xsl, which produces an RSS 0.92 feed. Please send suggestions/fixes to RSS-DEV.

Try it

XSL file:

XML data:

Base:

Channel:

Page:

Background

Following the Aug 2000 announcement of RSS-ala-RDF (thanks Ian, Rael, Danbri, and Guha!), and my notes on Semantic Web screen scraping (e.g. HyperRDF), I developed this little transformation that turns the W3C home page into RSS.

This isn't a really original idea; see XHTML-to-RSS Extractor service by Evic van der VList and DanBri; one of the RSS tools.

Frameworks and Applications

W3C has taken a "framework" approach to things like channels, packages and such. We had the CDF submission in March '97, OSD and DRP in Aug '97. We held a Sep '97 push workshop and ended up pursuing RDF, a general framework for resource description and metadata, rather than particular vocabularies.

The idea behind this sort of generalized approach is that it adds just a little cost/complexity to each of the applications, but in return, each application is part of a richer whole. Convincing the developers of the first few applications when they can't see the benefits is a challenge. (See also: the HTTP extension framework [@@link], XML namespaces, etc.)

But imagine what the Web would be like if you needed a different client for search services, online documentation, community forums, and online banking. Actually, you don't have to imagine, you can probably just remember: you have used pre-web online information services, haven't you? ;-)

RSS is a very successful metadata application, and I am enthusiastic about the move to (re-)align it with RDF. In this little application alone, I'm using RDF to leverage the RSS vocabulary and the Dublin Core vocabulary in the same description of our site.

Related Work, Future Directions

Dom and Susan have done the bulk of the work of applying this technology in W3C's operations. See especially Automating the publication of Technical Reports. It would be nice to do software releases ala DOAP.

The Grounding link relationships and classes section in HyperRDF is some earlier, related work. GRDDL is a generalization of these techniques.


Dan Connolly, Aug 2000
$Revision: 1.26 $ of $Date: 2005/05/20 20:04:50 $ by $Author: connolly $

$Log: Overview.html,v $
Revision 1.26  2005/05/20 20:04:50  connolly
item title can be h2 or h3

Revision 1.25  2004/11/05 22:13:27  connolly
restored meta anchor

Revision 1.24  2004/11/05 19:20:55  connolly
whitespace, linebreak

Revision 1.23  2004/11/05 10:41:53  dom
more subtle markup to identify "#" as a profile rather than the HTML document itself

Revision 1.22  2004/11/04 16:18:04  dom
really

Revision 1.21  2004/11/04 16:16:58  dom
proper GRDDL transformer link

Revision 1.20  2004/11/04 16:14:19  dom
uri simplification

Revision 1.19  2004/11/04 16:12:26  dom
getting rid of RDDL, making it valid XHTML Strict

Revision 1.18  2004/11/04 15:15:15  connolly
Update prompted by weekly markup checker:
 - revert title/h1 to be about one convention for site summaries,
   not all profiles at W3C
 - revise how-to text to emphasize how-to over background spec links
 - fold rel="meta" convention into the one profile
 - made the RSS 0.92 stuff into complete sentences
 - relativize more links
 - fold references inline
 - folded acks into background; updated related/future work
 - localize move RDDL ns decl; hope to remove it
 - us-ascii encoding
 - added CVS changelog

----------------------------
revision 1.17
date: 2004/10/19 16:03:04;  author: dom;  state: Exp;  lines: +7 -6
grddl-enabling
----------------------------
revision 1.16
date: 2003/09/01 17:40:11;  author: slesch;  state: Exp;  lines: +1 -1
tweak
----------------------------
revision 1.15
date: 2003/09/01 16:55:01;  author: slesch;  state: Exp;  lines: +168 -118
added meta link type
----------------------------
revision 1.14
date: 2001/05/31 17:24:11;  author: danbri;  state: Exp;  lines: +1 -2
removed newline wrap
----------------------------
revision 1.13
date: 2001/05/31 17:15:47;  author: danbri;  state: Exp;  lines: +15 -1
added information about the RSS 0.92 experiment
----------------------------
revision 1.12
date: 2000/11/17 16:31:31;  author: connolly;  state: Exp;  lines: +3 -4
changed the title to connote less endorsement
and to reflect the applicability of this
tool to any/all web sites.
----------------------------
revision 1.11
date: 2000/11/16 23:58:07;  author: danbri;  state: Exp;  lines: +7 -3
removed link to alternate version, added mention of the recent change
----------------------------
revision 1.10
date: 2000/11/16 22:53:59;  author: danbri;  state: Exp;  lines: +5 -2
added link to new XSLT file
to track changes in the RSS 1.0 proposal.
----------------------------
revision 1.9
date: 2000/09/16 06:16:03;  author: connolly;  state: Exp;  lines: +31 -21
(connolly) Changed through Jigsaw.
----------------------------
revision 1.8
date: 2000/08/31 22:10:35;  author: connolly;  state: Exp;  lines: +2 -2
fixed channel param
----------------------------
revision 1.7
date: 2000/08/31 22:01:47;  author: connolly;  state: Exp;  lines: +19 -1
added form
----------------------------
...
----------------------------
revision 1.1
date: 2000/08/22 21:59:36;  author: connolly;  state: Exp;
Overview.html