ISSUE-19 (declarations-p): REPORTED: Resolve whether to include declarations

ISSUE-19 (declarations-p): REPORTED: Resolve whether to include declarations 

http://www.w3.org/2007/OWL/tracker/issues/

Raised by: Alan Ruttenberg
On product: 

Reported by alanruttenberg, Aug 08, 2007
Pro: 
  Allows for detecting some kinds of errors
  Allows expression of intention in otherwise ambiguous cases
Con:
  Some duplication of information
  Uncertain impact
  Not a lot of experience with it

(there may be others)

See:
http://lists.w3.org/Archives/Public/public-owl-dev/2007JanMar/thread.html#msg80
http://lists.w3.org/Archives/Public/public-owl-dev/2007JulSep/thread.html#msg27

Comment 1 by bparsia, Aug 13, 2007
It should be noted that declarations *exist*, in a sense, in OWL DL. Use typing and declarations are handled by  a mix of type triples and inference from syntactic position. Type triples are overloaded to be a kind of  declaration as well as a typing mechanism.

The second pro conflates issues concerning intention and issues concerning ambiguity. Ambiguity is generally  handled, in OWL 1.1, by manifest typing, e.g., "ObjectPropertyDomain" (vs. plain rdfs:domain in OWL). Now,  arguably, this signals your intention that the property in question is an object property (and the object of the  assertion is a class). But there are cases where nothing is ambiguous, but the author didn't intend for a  property to be punned as a class (current declarations don't enforce this bit). Thus, some form of declaration  could express this bit of authorial intent.

Comment 2 by bparsia, Aug 13, 2007
See also:

   http://code.google.com/p/owl1-1/issues/detail?id=37

But this deals more with the syntax, afaict.

Received on Thursday, 25 October 2007 04:34:37 UTC