Issue 10843: Annex D.2, OWL Full (odm-ftf) Source: NIST (Dr. Conrad Bock, conrad.bock(at)nist.gov) Nature: Clarification Severity: Significant Summary: Annex D.2, OWL Full. In Annex D.2, under Figure D1, the second sentence says that OWL Full must be used to subclass OWL class. Why can't OWL class be subclasses in OWL Lite or DL? All instances of subclasses of OWL:Class are also OWL classes, and presumably wouldn't violate the constraints of OWL Lite or DL just because of the subclassing. Resolution: Revised Text: Actions taken: March 30, 2007: received issue January 19, 2009: closed issue; Closed; No Change Discussion: Discussion: OWL DL and OWL Lite prohibit "messing with the vocabulary". This is intended to prevent users from changing the semantics of the language. OWL Class is an element of this vocabulary. When we asked Description Logic experts about this issue, they reported that the OWL DL semantics prohibits subclassing OWL Class. This prohibition also holds for OWL Lite which is a syntactic subset of OWL DL. Thus the text at issue is correct, subclassing OWL Class puts one into OWL Full. Disposition: Closed, no change End of Annotations:===== m: webmaster@omg.org Date: 30 Mar 2007 00:47:57 -0500 To: Subject: Issue/Bug Report -------------------------------------------------------------------------------- Name: Conrad Bock Company: NIST mailFrom: conrad.bock@nist.giv Notification: No Specification: Ontology Definition Metamodel Section: FormalNumber: ptc/06-10-11 Version: RevisionDate: Page: Nature: Clarification Severity: Significant HTTP User Agent: Mozilla/5.0 (Windows; U; Windows NT 5.1; en-US; rv:1.7.2) Gecko/20040804 Netscape/7.2 (ax) Description Annex D.2, OWL Full. In Annex D.2, under Figure D1, the second sentence says that OWL Full must be used to subclass OWL class. Why can't OWL class be subclasses in OWL Lite or DL? All instances of subclasses of OWL:Class are also OWL classes, and presumably wouldn't violate the constraints of OWL Lite or DL just because of the subclassing. From: ewallace@cme.nist.gov Date: Sun, 26 Aug 2007 12:39:15 -0400 (EDT) To: odm-ftf@omg.org X-Sun-Charset: US-ASCII X-MailScanner-Information: Please contact postmaster@mel.nist.gov for more information X-MailScanner: Found to be clean X-MailScanner-SpamCheck: X-MailScanner-From: ewallace@mailhost.mel.nist.gov Subject: issue 10843 discussion X-Spam-Status: No Reply from a Description Logic expert on the question raised in issue 10843 as to whether subtyping owl:Class will put one in OWL Full. Summary: Bijan says that it will. -Evan ----- Begin Included Message ----- >From bparsia@cs.man.ac.uk Fri Aug 24 17:50:33 2007 Mime-Version: 1.0 (Apple Message framework v752.2) Cc: ian.horrocks@comlab.ox.ac.uk, sattler@cs.man.ac.uk Content-Transfer-Encoding: 7bit From: Bijan Parsia Subject: Re: OWL DL sublanguage syntax restrictions Date: Fri, 24 Aug 2007 22:50:05 +0100 To: ewallace@cme.nist.gov On Aug 24, 2007, at 9:10 PM, ewallace@cme.nist.gov wrote: > Hello, > > I am looking for sage advice from some OWL DL language lawyers (as > Alan Rector sometimes refers to this role :). Quite simply I need > some clarification on one of the "Rules of Thumb for OWL DL > ontologies" included in the OWL Reference document. The rule in > question is "Don't mess with the vocabulary". I had originally > interpreted this rule rather broadly, And correctly. > but a colleague, Conrad Bock, > has stated that he has subtyped things in the OWL vocabulary and had > validator/ species detectors claim this was still OWL DL. Unfortunately, place not your faith too much in the validators. They are not the most tested bit of code out there and messing with the vocab happens seldom enough that bugs won't be well fleshed out. > While this > may not be redefining the language elements being so subtyped, It's forbidden. > it is a > form of metamodeling. I wouldn't have expected OWL DL reasoners to > understand the resulting extended vocabulary. If they do anything, they pun that vocabulary. Pellet does this with things like rdf:List. > As Bob Colomb > pointed out though, supporting this sort of extension would give OWL a > similar feature to the profiling capability of UML. UML stereotypes > provide the mechanism for subtyping UML language elements (its > vocabulary), but what they do is the same. > > This question has come up multiple times, but my reason for asking > today is to resolve an issue brought against the OMG Ontology > Definition Metamodel (ODM) specification. In particular, one > informative section claims that subclassing OWL:Class would put one in > OWL Full. Yes. > This is in dispute, and I haven't been able confirm either > claim. It's also important to clarify this for the normative OWL DL > metamodel so that we can include the proper OCL constraints. Here's the place to look in the spec: Notice that owl:Class is part of the disallowed vocabulary. The definition of separated vocabulary forbids any axiom or fact in abstract syntax form from mentioning the disallowed vocabulary. Thus, an axiom like Class(MyClass, parital, owl:Class) or SubClass(MyClass, owl:Class) is forbidden. Cheers, Bijan. Reply-To: From: "Conrad Bock" To: , Subject: RE: issue 10843 discussion Date: Mon, 27 Aug 2007 09:19:25 -0400 X-Mailer: Microsoft Office Outlook 11 Thread-Index: AcfoBte4AvRPHhnPRRWx4EybhlMdfwApYynw X-MailScanner-Information: Please contact postmaster@mel.nist.gov for more information X-MailScanner: Found to be clean X-MailScanner-SpamCheck: X-MailScanner-From: conrad.bock@nist.gov X-Spam-Status: No Evan, > Reply from a Description Logic expert on the question raised in > issue 10843 as to whether subtyping owl:Class will put one > in OWL Full. > > Summary: Bijan says that it will. I guess it's the letter of the law, but instances of subclasses conform to the superclasses, so I'm not sure how it would matter to the reasoners. It could be a case where the rules are more strict than necessary. It doesn't show up in Menzel's axiomization, for example. Conrad