[OEP and PORT] Another practical question

I hope the subject markers are ok - two prior related threads had these.

This question boils down to something like: when designing an ontology, 
is it better practice to maximise equivalences with existing ontologies, 
or to trade this interop for decidability?

A while ago I roughed out an ontology [1] to mirror the Atom syndication 
format [2] . Henry Story has recently been revising this, as he aims to 
use the ontology as the basis of the model in his weblogging tool 
(bloged [3]). He soon noticed the overlap with terms from FOAF and 
Dublin Core (the latter are laid out in all there glory at [4]) and so 
has proposed adding references to these to the ontology as OWL equivalences.

I'd hoped to keep the Atom/OWL ontology OWL DL (to enable the use of DL 
reasoners). Correct me if I'm wrong, but I believe these equivalences 
would force the ontology to OWL Full, as the mapped vocabularies use 
rdfs:Property etc. So what do I do?

This is clearly close to the "DL or Full?" question, but  the choice 
isn't as clear cut because of existing material.

If best practice happens to be "depends on what you want to do with it", 
fair enough, but if there's a way of getting the best out such a 
situation (or I'm working on false premises!), I'd like to hear it.

Cheers,
Danny.

[1] http://dannyayers.com/2004/02/atom/index.html
[2] http://www.intertwingly.net/wiki/pie/FrontPage
[3] https://bloged.dev.java.net/servlets/SummarizeList?listName=users
[4]  http://internetalchemy.org/2004/03/theNucleusOfAtom

-- 

Raw
http://dannyayers.com

Received on Wednesday, 12 May 2004 05:18:18 UTC