Bernard Vatant wrote:
Hi all
Agreed with Dan and all others saying we have to welcome Google's
move. But nevertheless, I take the risk to include myself in the 1000
defined below ... :-)
I suppose pages such as [1] with indications for webmasters are likely
to be more read by webmasters than RDFa specs themselves or linked
data best pratcices documents. So, is this page making correctly the
case for linked data? For structured semantic data, yes, and nevermind
the vocabulary.
But for linked data, well, not much. Linked data ate about
relationships, and unfortunately the only example given in this page
defining a relation between resources using "about" is "for the
structured data geeks out there" ... and can be misleading for people
not aware of what LOD is about.
<div xmlns:v="http://rdf.data-vocabulary.org/"; typeof="v:Person">
<span property="v:name">John Smith</span>
<span rel="v:affiliation">
<span about="http://en.wikipedia.org/wiki/Acme_Corporation";
property="v:name">ACME</span>
</span>
...
</div>
So John Smith is affiliated to a wikipedia page. Whoever has the ear
of Google folks behind this could simply suggest to replace in this
example "http://en.wikipedia.org/wiki/Acme_Corporation"; by
"http://dbpedia.org/resource/Acme_Corporation";, explaining quickly the
difference.
Of course one can wonder if a fictional guy is better off being
affiliated with a fictional corporation than with a real web page.
That said, to follow-up with Dan's suggestion, would it be really
difficult e.g., for LOD html pages such as
http://dbpedia.org/page/Acme_Corporation to be RDFa-ized?
Bernard
Bernard,
Google will engage the Linked Data realm based on the natural effects of
the Web. Note, they crawl DBpedia like crazy (since inception). Also
remember, R. Guha has been with Google for eons, he is one of the
original creators of what we know today as RDF.
The hidden story here ultimately comes down to: what's making Google
decided to unveil what's been hidden behind their curtains for a while?
My guess: Wolfram Alpha and Yahoo! Search Monkey. Businesses don't like
opportunity costs, and they absolutely don't twiddle thumbs once it
becomes palpable :-)
Others: As I state repeatedly: History is a great teacher. History can
endow you with immense clarity (technically and/or commercially).
Kingsley
[1] http://google.com/support/webmasters/bin/answer.py?answer=146646
On 13/5/09 15:23, Kingsley Idehen wrote:
I desperately hope that you can see the Google is providing a huge
opportunity to showcase Linked Data meme value. Again, so what -- if
they don't use existing vocabularies? What matters is that they are
using RDFa to produce structured data, and that is simply huge!!!
Yeah, to be blunt, the last thing this situation needs right now is
having 1000 semantic web pedants descend, complaining that they're
not doing x, y or z right, that they don't "get it", that they're
copycatting yahoo, or whatever. This won't help anyone and would be
severely counterproductive.
What would help right now is having real and sizable sites expose
lots of RDFa HTML pages using FOAF, DOAP, SIOC, SKOS, CC etc. If
anyone has such information and is exposing it only in RDF/XML and
not RDFa, I'd suggest looking to make that change...
Dan
--
Regards,
Kingsley Idehen Weblog: http://www.openlinksw.com/blog/~kidehen
President & CEO
OpenLink Software Web: http://www.openlinksw.com
