John,
I read your message this morning and thought you made a really good point. Then I read the rest of my accumulated mail and saw how many people had already responded to
other parts of your post. This is the hazard of sleeping.
There are still two things I want to say:
<JM> My objective is not to lampoon. It is to say this: discussion of 'reuse' might productively distinguish between reusing class hierarchies, separately from reusing property hierarchies, separately from reusing
collections of individuals (and separately from reusing axioms).
That’s part of what I wanted to communicate when I asked about definitions of ontology reuse. And it’s even more complex, of course. Using a single IRI that someone else
has created is a form of ontology reuse. Reuse need not involve an owl:imports axiom. Given how quickly reasoners bog down as ontology complexity and KB size increase, a semantic web application designer must carefully weigh the tradeoffs between performance
and semantics. I know there’s already a considerable body of work on ontology modularization. With no disrespect intended to any researchers, may I say that it strikes me as premature: We don’t have enough experience with large interlinked ontologies and KBs
to assess the balance between expressiveness and performance. I have my own opinions but they are hypotheses, not theories. In other words, this is an area in which we can conduct controlled experiments. Let’s hope that’ll happen so we can really have some
evidence for what works. (Unlike in software reuse, I might add.)
Second point: Why can’t we get away from XML? I don’t care how many standards use it. It’s not suited to human comprehension. There are better concrete syntaxes for RDF
and OWL. If you have problems with ISO 15926, that’s one thing, but please don’t compound the problem by using the most verbose, special character-encumbered representation. The Turtle form is so much better:
:isClassOfApprovedIn-o-hasClassOfApprover
rdf:type owl:ObjectProperty ;
owl:propertyChainAxiom ( inverseRoles:isClassOfApprovedIn part2:hasClassOfApprover ) .
Whatever one may think about RDF and OWL, criticizing them for having an XML representation is like criticizing Fortran IV (my first programming language!) for accepting
both .LE. and <= as relational operators.
Steve Wartik
-----Original Message-----
From: ontology-summit-bounces@xxxxxxxxxxxxxxxx [mailto:ontology-summit-bounces@xxxxxxxxxxxxxxxx] On Behalf Of John McClure
Sent: Wednesday, January 29, 2014 2:47 AM
To: ontology-summit@xxxxxxxxxxxxxxxx
Subject: Re: [ontology-summit] [ReusableContent] Partitioning the problem
With all the talk about 15926 I just had to look at it.
<rdf:Description>
<rdfs:subPropertyOf
rdf:resource="#isClassOfApprovedIn-o-hasClassOfApprover"/>
<owl:propertyChain rdf:parseType="Collection"> <rdf:Description rdf:about="&inverseRoles;#isClassOfApprovedIn"/>
<rdf:Description rdf:about="&part2;#hasClassOfApprover"/>
</owl:propertyChain>
</rdf:Description>
Can Everyman understand this? Then it got really funny.
<owl:Class rdf:about="#RightNamespace">
<owl:disjointWith rdf:resource="#LeftNamespace"/> </owl:Class>
Hard to know if this is a political statement or not. Then, still feeling outsmarted in recent threads, I twisted my brain into a knot trying to understand
<owl:ObjectProperty rdf:about="#isClassOfSubclassIn"> and <owl:ObjectProperty rdf:ID="hasClassOfSubclass">
My objective is not to lampoon. It is to say this: discussion of 'reuse'
might productively distinguish between reusing class hierarchies, separately from reusing property hierarchies, separately from reusing collections of individuals (and separately from reusing axioms).
It seems unwise to treat ontologies as indivisible units when answering why reuse is not happening in general or, more to the point of this years communique, why ontologies are NOT being used (at all?) in Big Data analytics. But if the
dependencies are such that an ontology cannot be so divided, then maybe just maybe that's one reason why reuse is not occcuring?
/jmc
_________________________________________________________________
Msg Archives:
http://ontolog.cim3.net/forum/ontology-summit/
Subscribe/Config:
http://ontolog.cim3.net/mailman/listinfo/ontology-summit/
Unsubscribe:
mailto:ontology-summit-leave@xxxxxxxxxxxxxxxx
Community Files:
http://ontolog.cim3.net/file/work/OntologySummit2014/
Community Wiki:
http://ontolog.cim3.net/cgi-bin/wiki.pl?OntologySummit2014
Community Portal:
http://ontolog.cim3.net/wiki/