W3C home > Mailing lists > Public > public-rdf-dawg@w3.org > July to September 2005

Re: Prevelance of bNode-as-URI queries

From: Steve Harris <S.W.Harris@ecs.soton.ac.uk>
Date: Thu, 14 Jul 2005 12:04:15 +0100
To: DAWG public list <public-rdf-dawg@w3.org>
Message-ID: <20050714110415.GG27994@login.ecs.soton.ac.uk>

On Thu, Jul 14, 2005 at 06:59:40AM -0400, Bijan Parsia wrote:
> 
> On Jul 14, 2005, at 6:51 AM, Steve Harris wrote:
> 
> >I'm analysing the query logs of our main demo RDQL query server (we
> >allow members of the public to query and browse it) for a project.
> >
> >Out of interest I thought I'd find how many queries are handled where
> >people use the nasty <_:bnodeId> trick, turns out over the last 20 
> >days we
> >handled 55,000 of those queries per day, which is way more than I
> >expected, there in around 6% of all queries. There were queries on 
> >about
> >70,000 different bNodes over the 20 days, so its not one or two 
> >important
> >instances.
> 
> That's really interesting.
> 
> >That store does contain a lot of FOAFish data, so there are a lot of
> >bNodes in it, but it hints to me that this is a more important feature
> >than I thought.
> 
> I invite all to join me in my quest to rid FOAF and the world of the 
> believe that the inverse functional datatype on BNode trick is a good 
> idea. :)
> 
> Stand up and be URIed!

Amen brother. bNodes are a constant pain in the arse.

- Steve
Received on Thursday, 14 July 2005 11:04:48 GMT

This archive was generated by hypermail 2.3.1 : Tuesday, 26 March 2013 16:15:23 GMT