The Sun BabelFish Blog

naming disputes? No more!

Last weekend I spent some time drawing up a UML diagram of AtomRDF [1]. AtomRDF, written by David Powell, is a competitor to AtomOwl, the ontology I have been working on for a couple of years to model the Atom syndication spec as it evolved. David Powell and I have been engaged in a similar task but we have been mostly been working independently. I was trying to work out the problems I cared most about in my own way, and I am sure the same is true of David Powell. So in the end we ended up with two ontologies located in different name spaces. As both of these have been quite stable now, we have been looking at merging them. The beauty of RDF is that in the process of drawing up the AtomRDF UML diagram I could identify which of his names referred to the same things my names were referring to and just map most of the vocabulary from AtomRDF to AtomOWL, using sentences (triples) such as the following (see AtomOwl2RdfBridge.n3 attachement)

@prefix awol: ≤http://www.w3.org/2005/10/23/Atom#≥.
@prefix ardf: ≤http://djpowell.net/schemas/atomrdf/0.3/≥ .
@prefix owl:  ≤http://www.w3.org/2002/07/owl#≥ .

ardf:entryInstance owl:equivalentClass awol:version .
awol:entry owl:inverseOf ardf:containingFeed .

#
#Category
#

ardf:Category owl:equivalentClass awol:Category .
ardf:categoryTerm owl:equivalentProperty awol:category .
ardf:categoryScheme  owl:equivalentProperty awol:scheme .
ardf:categoryLabel owl:equivalentProperty awol:label .

#
# awol:Id classes does not exist in atomrdf
#

ardf:entryInstance owl:subPropertyOf  awol:version .
ardf:feedInstance owl:subPropertyOf  awol:version .
...

(( Now that is seems to me is a much better way dissolve meaningless debates about vocabularies than the method Dave Winer presented in his NerdTV appearance: to find the most stupid name that would fit. ))

Having found a good mapping for most of our concepts, we can now focus on those terms there seems to be real divergence on. There are some good things I can take from David Powell's ontology, and there are probably some good things he can take from mine.

Rdf brings us right to the substance of the work we are doing. As Robert Nozick argued in Invariances : The Structure of the Objective World by finding the invariances between theories or ways of looking at the world, we get closer to an objective view of it. Invariances are found by finding mappings from one theory to another. In our case by using the OWL vocabulary we can map our two ontologies together, and in the process feel a lot more comfortable that we have captured the concepts correctly. This is how we can come closer to an objective understanding of what Atom really is about.
There is more work to do, but we have come a long way in understanding each other now.

Notes

1. I found SWOOP very useful to put together a UML diagram of AtomRDF.

Posted at 05:43PM Mar 24, 2006 [permalink/trackback] by Henry Story in SemWeb  |