W3C home > Mailing lists > Public > public-owl-dev@w3.org > October to December 2007

Re: bags of axioms

From: Danny Ayers <danny.ayers@gmail.com>
Date: Wed, 31 Oct 2007 21:14:25 +0100
Message-ID: <1f2ed5cd0710311314t5b461161j1b1dc25ee1481687@mail.gmail.com>
To: "Matthew Pocock" <matthew.pocock@ncl.ac.uk>
Cc: public-owl-dev@w3.org

I'm only beginning to get up to speed on OWL 1.1, but a lot of what
you describe does sound in more familiar territory, so here's a few
cents worth -

On 31/10/2007, Matthew Pocock <matthew.pocock@ncl.ac.uk> wrote:

[snip]

I would suggest treating the material as linked data [1] rather than
hiding the messages in SOAP. This doesn't necessarily help with the
immediate problem, though describing the sets in a manner akin to
named graphs may suggest an approach that's consistent across both the
payload and the protocol.

Our system allows a certain amount of slop in what gets
> passed over the wire, within defined bounaries, so two identical requests
> against an identical service may legitimately return different sets of
> axioms, without the requester becomming confused or behaving differently.

That doesn't sound dissimilar to the usual WebArch notions of a
resource and its representations. My homepage is the same resource,
despite it being very variable slop. Why not give "these sets of
axioms" a URI..?

> Hence, the only thing that uniquely identifies the set of axioms is the set
> of axioms itself.

...but if that truly is the case, then your only option would appear
to be to use an opaque datatype, probably represented logically as a
functional property of something you can name as a resource,
physically as a block of XML.

> So, the things we are moving about:
>
> a) have no physical location, ever (they are generated by software, streamed,
> consumed by software)

Way too metaphysical for me ;-)

> b) have no logical URI that identifies them, or any other identifier smaller
> than themselves

...yet?

> c) are a sub-set of the axioms entailed by some other ontology, which does
> have a logical URI

Sounds like there may be a subsumption relation there somewhere.

> Is there some other top-level element other than Ontology that better fulfills
> these needs?

Offhand, I've now idea, but I see not reason not to define one.

> If not, is there some standard way to indicate to tools that the ontology has
> no physical or logical URI, but is a) a sub-set of the axioms in another
> ontology or b) the axioms entailed by another ontology or c) a sub-set of the
> axioms entailed by another ontology?

RDF uses bnodes where naming is difficult or inappropriate (and they I
believe are defined as existentials). I'm not sure under what
circumstances the OWL 1.1/RDF mapping holds, but it might be worth
exploring.

Cheers,
Danny.

[1] http://en.wikipedia.org/wiki/Linked_Data

-- 

http://dannyayers.com
Received on Wednesday, 31 October 2007 20:14:34 GMT

This archive was generated by hypermail 2.3.1 : Wednesday, 27 March 2013 09:32:55 GMT