RE: OWL 2 Profiles and Horn-SHIQ

Hi Ian,

 

Thank you for your feedback. I completely understand the working group's
concern that having too many profiles may end up confusing potential users.
Further, while we have clear internal use cases for Horn-SHIQ, given the
nature of our business it would be quite difficult for me to talk about them
in such a public setting. This said, thank you again for your feedback and
best of luck with the working group.

 

Cheers,

Chris

 

 

--

Christian Halaschek-Wiener, Ph.D.

Chief Technology Officer

Clados Management LLC

 

 

From: Ian Horrocks [mailto:horrocks@cs.man.ac.uk] 
Sent: Friday, July 18, 2008 5:57 AM
To: Christian Halaschek-Wiener
Cc: public-owl-comments@w3.org; OWL 1.1
Subject: Re: OWL 2 Profiles and Horn-SHIQ

 

Dear Christian,

 

When discussing fragments (now called profiles) the OWL Working Group
decided that having too many different profiles might be confusing to
potential users of OWL, and that  we should therefore try to limit the
number of profiles and only describe those for which (a) there is a clear
user requirement; (b) there is sufficient implementer support; (c) there are
significant distinguishing characteristics (e.g., computational or
implementational) w.r.t. other profiles. The WG is not aware of strong user
requirements or implementer support for Horn-SHIQ, and it's characteristics
seem to be quite similar to those of the OWL-R profile.

 

If you would like the WG to reconsider, could you provide additional
information about Horn-SHIQ that addresses the above points.

 

Regards,

Ian Horrocks

Co-Chair, W3C OWL Working Group

 

 

On 30 May 2008, at 18:14, Christian Halaschek-Wiener wrote:





Greetings,

 

I'm with an investment firm currently using OWL for a variety of purposes
internally within our company (some of you know me previously from my time
spent in Jim's Mindswap research group at UMD). We're very interested in the
work the group is doing on defining tractable subsets/profiles of OWL 2, as
performance guarantees are very important for our usage. I recently happened
to notice that one of the original member submissions related to tractable
fragments [1] included Horn-SHIQ. While being syntactically restricted and
having intractable taxonomic complexity, this fragment is interesting to us
because it provides modeling constructs not possible with the current
profiles defined in [2] (e.g., simultaneous usage of existential
quantification, inverse & transitive roles, etc.) and still has tractable
data complexity. I was wondering if there were specific reasons why this
profile is not currently included in the latest working draft? Thanks for
any insight or clarity the group can provide.

 

Cheers,

Chris

 

[1] http://www.w3.org/Submission/owl11-tractable/

[2] http://www.w3.org/TR/owl2-profiles/

 

--

Christian Halaschek-Wiener, Ph.D.

Chief Technology Officer

Clados Management LLC