Re: Survey of RDF data on the Web

Hi Seth,

> > From: Seth Russell [mailto:seth@robustai.net]
> > Sent: Thursday, August 15, 2002 9:49 AM
> > To: Andreas Eberhart; www-rdf-interest@w3.org
> > Subject: Re: Survey of RDF data on the Web
> >
> > From: "Andreas Eberhart" <andreas.eberhart@i-u.de>
> >
> > > [2] http://www.i-u.de/schools/eberhart/rdf/
> >
> > If you gonna publish data to a RDF interest group, you might get a better
> > response if the data was in RDF.
> >
> > ... just my humble opinion
[...]
> So?  ... read the text of the data in your sailor agent ...


While it would be very interesting to see an RDF/XML version of this
survey shared with the RDF Interest Group, I'd appreciate it if
listmembers took care to balance feature requests with encouragement.
There's no such thing as a free survey! Designing and implementing an
RDF/XML dump of this data would be good. But it's work; timeconsuming
work. So please don't complain when people share their work in this forum.
In email, enthusiasm ('hey, what a cool survey, would be even more useful
if it was available in RDF')  can easily come across as complaint ('this
should be in RDF! Plain text data not appropriate here!'). I don't do much
chairing on this list, but I do try to encourage folk to make this a list
where people are happy to share their work, and their works-in-progress.

So... I disagree with the suggestion in
http://lists.w3.org/Archives/Public/www-rdf-interest/2002Aug/0074.html
that people should wait until they've worked out an RDF view of their data
before sharing findings with this group. Publish early and often!

All that said, I for one would really appreciate an HTML/XHTML version of
the report[1], or even a PDF version. I'm often browsing/reading from
machines without Postscript installed, and it is also harder to quote from
and reference subsections of postscript documents.

BTW on the topic of stats from RDF harvesters, I did make a quick attempt at a
dump of such info from my 'scutter' harvester for FOAF data. Sample log at
[2], more on FOAF at [3]. Software distrib is pre alpha, but can be found
c/o [4]. The current log (generated by SQL query from a Postgres DB) lists
the number of occurances of each class and property encountered by the
harvester. I'd like to extend this to keep stats of class-contextualised
property occurances, for eg not just the number of dc:title-s, but the
number of dc:title properties applied to any xyz:AudioWork. Hypothesis
being that such data could be used for query routing and service
aggregation across multiple harvester installations that had different
traversal and indexing policies...

cheers,

Dan


[1] http://www.i-u.de/schools/eberhart/rdf/rdf-survey.ps
[2] http://rdfweb.org/2002/foaf/scutter/log/vocabstats-2002-7-15.log
[3] http://rdfweb.org/foaf/
    http://xmlns.com/foaf/0.1/
    http://www-106.ibm.com/developerworks/xml/library/x-foaf.html
[4] http://rdfweb.org/2002/foaf/software/ (a GPL'd Ruby RDF harvester)
    http://rdfweb.org/2002/foaf/scutter/ (unpacked distribution of same)
    http://rdfweb.org/2002/foaf/scutter/doc/scutter.html (quick overview)

Received on Thursday, 15 August 2002 12:46:34 UTC