Ian,

I had already noted that no HP people attended the post-ISWC workshop.
I am glad you receive this with interest. I am in no position to speak
for the workshop - Peter Patel-Schneider is probably the best person to
talk to on this. The most important information are collected on the
mailing list (linked to below). This work has no W3C status indeed, and
it was unclear how to formally go ahead. There was a wide spread
agreement though that a "1.1" with low-hanging fruits should follow to
1.0 within a year from now, while more complex changes would need to
wait for a long-term "2.0" spec. No matter how great OWL 1.0 is, it
clearly lacks some crucial features (such as numeric datatype range
constraints) to achieve wider industry adoption.

I fully agree that Jena should not simply add these features without
making clear that we are talking about pre-beta stuff. On the other
hand side, any evolving standard requires a reference implementation so
that we (as a community) can detect pitfalls based on practical
experience. I guess given the very important role of Jena in the
evolution of OWL 1.0, it would be beneficial for all if Jena would again
take the lead :)

But I suggest anyone who is interested in this discussion should
subscribe to the new OWL mailing list, and continue the discussion on
that list.

Holger


Ian Dickinson wrote:
> Hi Holger,
>
>>are there plans to support Qualified Cardinality Restrictions (QCRs) in
>>the Jena OWL Profiles? As you certainly know, the OWL 1.1 specification
>>will contain QCRs [1], and it is very likely that the chosen syntax will
>>simply use an additional property owl:valuesFrom to indicate the class
>>filler. This syntax is already supported by Racer, Protege and as far
>>as I know also by other tools.
>>
>>While you will certainly respond that the group is waiting for 1.1 to be
>>released, it would be a major delay for a feature that seems rather
>>trivial to implement given that Jena already has this support in its
>>DAML+OIL profile.
>
> Actually, I'm going to respond by saying that I didn't actually know
> that an OWL 1.1 was in the works. As you say, it would be pretty
> straightforward to add QCR's to the OWL profile, less so for the DIG
> interface. Some of the other features seem a bit more work, especially
> for the reasoner implementors.
>
> Is there more to OWL 1.1 than Peter's list of wished-for features?
> Which WG is going to be releasing it? I'm loathe to put non-standard
> features into Jena's OWL profile. As an interim step, I suppose we
> could add a new OWL 1.1 profile, with a warning in large unfriendly
> letters that it's just a proposal. But before even doing this I'd like
> to understand where OWL 1.1 is coming from, and where it's going.
>
> Thanks for bringing it up though. Interesting.
> Ian
>
>
>
>
> Yahoo! Groups Links
>
>
>
>
>
>