On 6/12/08 6:53 PM, "Peter Ansell" <> wrote:
> 2008/6/13 Dan Brickley <>:
>> Alan Ruttenberg wrote:
>>> It's OWL, but not OWL-DL.
>>> I would very much like there to be an OWL-DL version too, or at least to
>>> factor it into two components - an OWL-DL portion, and a set of further
>>> axioms that are imported by OWL full users.
>> I've wondered about how best to do this: are there any discovery conventions
>> for finding an OWL DL flavour of a vocabulary which otherwise also has OWL
>> Full variants at the namespace URI? Or vice-versa? Could eg. editors read
>> HTML or RDF/XML from the namespace URI, poke around and find the URL to a
>> pure DL subset?
> Not as far as I know. I was chatting to Alan about this recently but
> the best I could suggest was an improvement in OWL 2.0 where people
> specify profiles (in RDF) within ontologies which people can choose
> when they wish to interpret the ontology. I know that W3C has been
> against the use of XRI's, because they have no clear benefit, but in
> the case of being able to specify the profile "flavour" of an ontology
> that you want to use, it would be beneficial to be able to say it
> without changing the URI, something with XRI's apparently provide.

As you have observed, the clear benefit of XRIs is their stability--a
characteristic which derives from them being "names" for things and not mere
"locations" of things. For example, I might put my wallet on my dresser or I
might put it on my desk (different locations), but it is always my wallet.
The trouble with ordinary URIs is that is
different from There is no separation of
concerns: naming vs. location. They are mixed together.

So for the W3C to say that "there's no clear benefit" while ,
simultaneously works on (re-)inventing its own durable identifiers (as is
underway in the past couple of years with Linked Data Hash URIs and 303
URIs, appears a bit disingenuous. 

Received on Friday, 13 June 2008 10:34:59 UTC