RE: RDF query test cases?

Hi Danny,

This sounds interesting, and is a limitation that irritates me too.

The various solutions we have mulled over are:
1). Simply putting the categories in a taxonomy is a good start (they're not
currently in MT)
2). Explicitly share a taxonomy (OK for small groups).  
3). Allow lightweight mappings between related taxonomies, using something
like an RDF version of XFML [1]

We're likely to adopt (1) and (2) for SWAD-E demonstrator 12.1, with (3) as
an optional extension
Your wordnet approach is another interesting angle, let us know how it goes!

Cheers

Steve

[1] http://xfml.org/  

-----Original Message-----
From: Danny Ayers [mailto:danny666@virgilio.it]
Sent: 21 January 2003 10:08
To: Jeen Broekstra; Libby Miller
Cc: www-rdf-rules; Steve Cayzer
Subject: RE: RDF query testcases?


This thread has provided the answer to a problem that I've been mulling over
for a few weeks ;-)

Some blogging tools (such as Movable Type) allow the user to categorise
their posts, but the categories used are totally arbitrary, just strings
decided by the user. To be able to index across systems, some sharing of
taxonomies would be needed. It occurred to me that a lookup of something
like Wordnet would allow the mapping of Cats (in Danny's blog) to
http://whatever/worndet#Cat and thence to Cats (in Libby's blog). So it
looks like pretty much any of these query tools would be up to the job. The
next stages are I suppose setting the 'dictionary' up as a service, then
implementing a user-transparent interface.

Cheers,
Danny.


-----------

http://dannyayers.com

"The lyf so short, the craft so long to lerne." - Chaucer



>-----Original Message-----
>From: www-rdf-rules-request@w3.org
>[mailto:www-rdf-rules-request@w3.org]On Behalf Of Jeen Broekstra
>Sent: 21 January 2003 10:27
>To: Libby Miller
>Cc: www-rdf-rules
>Subject: Re: RDF query testcases?
>
>
>
>Libby Miller wrote:
>
>> this is great, thanks Jeen.
>>
>> quick question: Sesame supports RDF schema right? so if I did a query
>> in RDQL over Sesame over an ontology like Wordnet (e.g.
>> http://xmlns.com/wordnet/1.6/Person) would I get both Person and
>> Life_form as the classes of an instance of Person?
>
>Yes.
>
>My RDQL is somewhat shaky, but the query would be something like:
>
>SELECT ?c
>WHERE (?p, <rdf:type>, <wn:Person>),
>       (?p, <rdf:type>, ?c)
>
>Right?
>
> > What if I did the same query in RQL?
>
>The same. In Sesame, the deductive closure is computed independently of
>the query module, so both RQL and RDQL could retrieve all these answers.
>
>The RQL query in this case would be something like:
>
>SELECT typeOf( p )
>FROM   wn:Person { p }
>
>The difference is in the fact that RQL can explicitly express certain
>types of schema semantics in the query, making it possible to express
>queries about the schema more easily, and sometimes go beyond what's
>expressible in an RDF-only QL.
>
>A simple example of this is direct subclass relations (A is a direct
>subclass of B iff there is no C: A < C < B): this is a relation that
>would be rather awkward to express in RDQL, but RQL has a special
>language feature for it.
>
>> I'm asking this because I was asked to recommend a tool that could be
>> used for querying a fairly simple, heirarchical ontology (actually
>> more like a thesaurus - MeSH,
>> http://www.nlm.nih.gov/mesh/meshhome.html), such that you could, say,
>> ask for all research groups with 'abdomen' or anything above it in the
>> heirarchy as their descriptive keyword. (I'm now also wondering whether
>> things below it in the heirarchy might be more useful...)
>>
>> Does that make any sense?
>
>This would be fairly straightforward to express, I imagine, so yes :)
>
>Best regards,
>
>Jeen
>--
>jeen.broekstra@aidministrator.nl
>aidministrator nederland bv - http://www.aidministrator.nl/
>julianaplein 14b, 3817 cs amersfoort, the netherlands
>tel. +31-(0)33-4659987, fax. +31-(0)33-4659987
>

Received on Tuesday, 21 January 2003 05:50:50 UTC