W3C home > Mailing lists > Public > www-rdf-interest@w3.org > August 2002

Re: Google Result in RDF

From: Alexander Löser <aloeser@cs.tu-berlin.de>
Date: Thu, 08 Aug 2002 17:31:23 +0200
Message-ID: <3D528ECB.A19D76D0@cs.tu-berlin.de>
To: Seth Ladd <seth@brivo.net>
CC: www-rdf-interest@w3.org

Seth Ladd wrote:

> Alexander,
> > Ok. Now a relatedInformationURI Property is aviable. You can use this Property
> > to initiate an new RooDolF search.
> That's wonderful!

Thank you.

> Of course, the relatedInformationPresent property
> isn't needed anymore.

Nope. I will still integrate this property, because some URLS dont have related information present.
And I would like to model this.

> To be a bit more RDF correct, the
> relatedInformation should be a property pointing to a resource (instead
> of a literal.
> <RooRolF:isRelatedTo
> rdf:resource="http://nutria.cs.tu-berlin.de/roodolf/RoodolfSearchResult.jsp?&amp;query=related..."/>
> This is a hard problem because the resource we're talking about (a web
> page somewhere on the web) isn't really related to the above URI.  It's
> related to items inside the above URI.  So how do we say "the URI X is
> related to the resources inside URI Y." where X is a search result and Y
> is the resource of the related information page?

Exact. It is very hard. Maybe other "Members of the List" can help us?!

> In fact, this is a very hard problem.  Google search results return URIs
> and information about those URIs in a Google context.  That is, the
> cachedSize property is really a property of the cached page at google,
> not the URI of the resource itself.
> Maybe a better way to model this would be:
> <rdf:Description rdf:about="http://www.example.com/some/real/uri/on/web">
>     <RooDolF:hasGoogleCache>
>      <rdf:Description>
>        <RooDolF:cachedImage
> rdf:resource="http://www.google.com/cache/to/page"/>
>        <RooDolF:cachedSize>24k</RooDolF:cachedSize>
>      </rdf:Description>
>     </RooDolF:hasGoogleCache>
> </rdf:Description>

Thank you, I don't know, if we should model the Google Cache. But maybe you are right. I'm relative new
to the RDF stuff.
In my opinion what we sould model is:
There is a Relation between Resource x and y.
This Relation is based in the Author: GoogleRelatedAlgorithm and has existed at Date xyz.

Maybe other experienced users could help here?

> > I don't know how to realize  such an query string with tomcat. Any Idea?
> >
> > The URI for the search is created already but for the adequat Google Search.
> > So your Property hasGoogleResults
> > can be described by adding the google URL.
> > Do you belive it would be interesting to create a Roodolf search Url Property?
> That's an excellent point.  I think I got confused as to what you are
> describing in your RDF.
> Really cool project,

Thank you again. Will you use the interface in your projects?  At the moment we are working  on a
similar interface for dmoz.
Please see also (draft) http://nutria.cs.tu-berlin.de/dmozrdf/index.html
Searches include at the moment some (adult) pages because we did not have enough space on our disc (1.5
GB is needed)

> Seth



  Alexander Löser
  Technische Universitaet Berlin
  Fakultaet IV - CIS
  bmb+f-Projekt: "NewEconomy"
  "Neue Medien in der Bildung"

  email: aloeser@cs.tu-berlin.de
  office: +49- 30-314-25555
  fax   : +49- 30-314-21601
Received on Thursday, 8 August 2002 11:35:33 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 15:07:42 UTC