- From: Rob Shearer <Rob.Shearer@networkinference.com>
- Date: Thu, 6 May 2004 16:12:04 -0700
- To: <kendall@monkeyfist.com>, "Seaborne, Andy" <andy.seaborne@hp.com>
- Cc: <public-rdf-dawg@w3.org>
> I'm a cell phone. I want to look at the RSS 1.0 feeds of all the > people in my FOAF file in order to construct a list of email > addresses. I want to submit one query to my DAWG service and get back > a list of email addresses (in some format... unspecified here). I want > to submit *one* query because my service is often unreliable and > always expensive. > > So, instead of requesting one query n times, I want to request > one. Something like: > > SELECT... > WHERE... > USING... > IN > http://foo/foaf.rdf > http://bar/foaf.rdf > http://blaz/foaf.rdf > > That's what I called aggregated query; but note that it doesn't > require the multiple graphs to be remote. I think there are other use > cases where they are local. Just what server are you asking to perform this query for you? What on earth is stopping some independent application from simply aggregating these graphs for you? It seems to me that such an application is quite a different beast from a general query answering service. I can completely understand a standalone RDF repository that lets you ask about the RDF it contains (because it wishes to offer knowledge of its contents). It seems far less likely that in order to offer an RDF query service that same repository needs to be willing to go and aggregate data for you because the query language says this is part of performing a query. Again, how you form your query graph is your problem. How you get data out of it is a problem for the query language (and possibly protocol).
Received on Thursday, 6 May 2004 19:13:39 UTC