W3C home > Mailing lists > Public > www-rdf-interest@w3.org > August 2004

Re: RDF graph merging question

From: Damian Steer <damian.steer@hp.com>
Date: Wed, 18 Aug 2004 00:19:09 +0100
Message-Id: <D781F3CE-F0A3-11D8-A4B6-000D932B9016@hp.com>
Cc: www-rdf-interest@w3.org
To: algermissen@acm.org

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1


On 17 Aug 2004, at 19:41, Jan Algermissen wrote:

> Example: Suppose two computers are connected and we represent this as
>
>   foo:host1 bar:connectedTo foo:host2
>
> Now, in RDF graph A the triple is reified to attach the information
>
>   foo:conn-host1-host2 baz:connectionType foo:ethernet to it
>
> (the connection is an ethernet connection)

I don't think this is right. That says a statement has a connection 
type 'ethernet', which is odd (to say the least).

What you want is an n-ary predicate, something like connection(host1, 
host2, type). I'd probably model it as

_:connection rdf:type foo;Connection .
_:connection foo:host _:host1 .
_:connection foo:host _:host2 .

(best practices have a document about this general problem [1])

OWL can tighten that a little, eg only two foo:hosts per connection.

> Now in some strore B, the same triple exists and is reified to attach
> a cable number:
>
>   foo:connection-123 baz:cableNumber "XY-T-5665"
>
>
> An RDF store (at least one that does not provide some non-standard
> extension) cannot by itself provide me with the information that
> the connection is of type ethernet and has cable number "XY-T-5665",
> right?

So we want to say 'if two connections have the same hosts then they are 
the same connection', yes? I don't think OWL can do that, but I may be 
wrong (suppose foo:host pointed to a collection of hosts instead, and 
was an IFP - does that work?). Property restrictions might do the 
trick, I guess, if you wanted to talk about classes of connections.

Going beyond OWL the rules engines can do this fairly simply, although 
I don't know if you count them as 'standard'. They're pretty 
widespread, at least.

> In case it seemed so: my intention was not to point out a conflict, but
> I am evaluating several data models for their suitability for data 
> integration
> and the issues above are my primary interest.

I hope this helps a little. More knowledgeable people will now correct 
me :-)

Damian

[1] http://www.w3.org/TR/swbp-n-aryRelations/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.2.4 (Darwin)

iD8DBQFBIpJ0AyLCB+mTtykRAqYSAKD0J39WvQ10KAr8d6GMbt2C6Px1FACggof2
YfbtciO7ySa5aixoYXXoJnA=
=NBz8
-----END PGP SIGNATURE-----
Received on Tuesday, 17 August 2004 23:19:51 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:14:57 UTC