Yeh, Yaron had the same question when we discussed it and it's absolutely true, that linking to RDF version of the page is clearly more functional, but there is an issue of those two documents (HTML page and RDF document) being different sources of info opening potential door for spammers and it looks like search engines like Yahoo who's implementing semantic web support are using single-source approach to try to limit this and that's why RDFa/eRDF are the only two approaches they support in current Search Monkey documentation.
In any case, it's better to have various means of embedding semantic data since we don't know what will be the prevalent approach for "smart agents" to gather this data.
Sergey Chernyshev wrote:Heh, I brought that up years ago when I first heard of RDF/a (below). I
> I'm thinking about different forms of RDF publishing that is being
> adopted by various data consumers and RDFa (and possible eRDF) seem to
> be quite interesting - Yahoo hopes to support it in SearchMonkey
> eventually and RDFa usecases at W3 even mention it in connection with
> SMW: http://www.w3.org/TR/2007/WD-xhtml-rdfa-scenarios-20070330/#use-case-6
> The question to Markus and Denny, are there any plans for this?
think I filed an enhancement bug but I can't find it now.
However, I wonder how useful RDFa in SMW is. Every SMW page explicitly
identifies that its facts are available as rdf+xml in the official
<link rel="alternate" type="application/rdf+xml" title="Berlin"
So is it only tools that ignore this and just look for RDFa in HTML that
benefit? Are there lots of such tools? The bookmarklet I mention below
(hah, it's been in my browser toolbar for over two years and I haven't
used it for 18 months) would highlight any RDFa statements in HTML.
That might be useful but a) SMW shows all its statements in the factbox,
and b) many SMW pages hide their statements (with the "| " alternate text).
I'm not arguing against it, just wondering.
= = = = = = = early message = = = = = = =
Subject: Re: [Semediawiki-user] Do (could) types show up in the HTML?
Date: Sun, 26 Mar 2006 22:51:12 -0800
Summary: consider RDF/A for this.
Rich Morin <email@example.com> wrote a while ago:
> I've looked at the source code for the page
> looking for any encoding of the type information in
> the generated links. I haven't found any, but I may
> not be looking in the right places, way, etc.
> In any case, I would like to promote the idea of
> making this information "visible" to web browsers,
> web crawlers, etc.
Denny Vrandecic <firstname.lastname@example.org> responded:
> Nice idea. Is there already a syntax for that, like
> <a href="target-url" type="link-type-url">shown text</a>
Yes. I think RDF/A , "syntax for expressing RDF metadata within XHTML",
is designed to do this. You use the rel= attribute of <a href> with a
value from a particular vocabulary's namespace. The cool thing about
RDF/A is it also provides a way to denote property values using <span
To misquote the fine http://wiki.ontoworld.org/index.php/San_Diego page,
It is the county seat of San Diego County,
with a population of 1,223,400.
I'm guessing its XHTML with type information encoded in RDF/A would be
It is the county seat of
rel="smw:_Relation-3AIs_county_seat_of">San Diego County</a>,
with a population of
One of the people on the "RDF in XHTML Taskforce" has a basic
bookmarklet to highlight RDF/A triples in HTML documents,
It would seem pretty feasible to add the <span property=> tag in the
SMWParseAttributesCallback() hook. Adding the rel= attribute to
relation hyperlinks is harder as SMW currently relies on MediaWiki's
default [[wiki link]] processing after it's parsed the relations.
The RDF/A primer
http://www.w3.org/TR/2006/WD-xhtml-rdfa-primer-20060310/ is pretty
approachable compared with other W3 documents (until it gets into URIs
and CURIEs :-).
Check out the new SourceForge.net Marketplace.
It's the best place to buy or sell services for
just about anything Open Source.
Semediawiki-devel mailing list