RE: A question on the vocabulary for 'persons' - ACL level of granularity?

--Dan,

> Yup. It's one thing to define a fresh new vocabulary 
> uncluttered with tributes to our Internet forefathers. It's 
> quite another to populate it with machine readable critiques 
> of sibling ontologies that hinder simultaneous usage.  I 
> guess you *could* write OWL that says something like "there 
> are no things that have any value for a foaf:geekcode 
> property". Or something slightly more conservative, "if it's 
> got one of those properties, it's not in the class 
> my:Patient!". But such idioms would serve little practical 
> purpose, and would in fact be usage/acknowledge of a 
> vocabulary construct ('geekcode') which the critiquer 
> considers incoherent/pointless/foolish. In RDF/OWL it is hard 
> (and counter-productive) for one ontology to try to say 
> "don't use that one over there, it's crap!". Better to vote 
> with one's feet, and to allow users and other apps to 
> pick-and-mix as they themselves prefer.

The foaf:geekcode example is a figure of speech especially with Chris'
objection to foaf. I certainly wouldn't expect that Chris or anyonelse in
that matter would in fact make some statement to recommend not to use
certain (or even part of an) ontology. But in science, there might be
situiation where two competing and contradicting theories coexists. Then
sometimes the "uncareful/unnecessary" statement may render an ontology
unsharable in the long run. 

And I absolutely agree with you that "Better to vote with one's feet".  I
have always thought that the best ontology those that are most shared, but
not necessarily those that are right or complete.  So, when designing an
ontology, our ultimate goal is to improve its sharing and reuse by others.
And careful modulization would certainly improve an ontology's sharing and
reuse.    

> >> Another remark, which may be too obvious to be worth 
> making, but here
> >> goes: You can use a namespace, and thus the symbols from 
> an ontology, 
> >> without importing it.  In some cases, one does this just 
> to declare 
> >> that you want to use that symbol to avoid making up one of 
> your own; 
> >> and you don't need the axioms that formally constrain the symbol's 
> >> meaning.  In other cases, there may be only a few such axioms, and 
> >> you can simply copy them.  I don't know if this is a good idea.  
> >> We're getting into a whole mess of hard questions about version 
> >> control, partial importing of ontologies, etc. etc. that I 
> wish I had 
> >> answers to.
> > 
> > Do you mean just use the URI without importing it? If so, I am not 
> > sure how it will work?  One of the neat features of the web is its 
> > loosely coupled nature.  But you need to follow your nose 
> to know more about the resource.
> > Without "importing", i.e., to fetch the resource 
> description from the 
> > namespace, what is the use of it?  For instance, if given a dubline 
> > core URI http://purl.org/dc/elements/1.1/creator, without following 
> > the URI, I won't even know how I should label it.
> 
> Related problem: If I write, "my:title rdfs:subPropertyOf 
> dc:title", ...
> am I dragging all instance data that uses my:title into the 
> commitments of the dc: namespace? (including being owl Full, 
> making DL reasoners bluescreen, etc)?  It would be nice to 
> have some conventions/standards for making low-cost re-use / 
> mappings of that kind. But right now we're in a world where 
> DL-compatible variants of the DC (and FOAF) namespaces 
> circulate in email, CVS, etc form, since such lightweight 
> re-use doesn't fit with the approach adopted in various OWL systems.

Yes, I agree.  There should be a clear specification regarding if a direct
reference should triggers an HTTP GET. I actually don't see any reason for
not doing so. And how it is related to owl:import, rdfs:isDefinedBy etc.

Xiaoshu

Received on Monday, 18 September 2006 03:56:11 UTC