Re: Possible to reduce OWL Full ontology to DL ontology automatically?

Jeremy Carroll wrotes:

> 
> Phil Dawes wrote:
> 
> > Jeremy Carroll writes:
> >  > 
> >  > Phil Dawes wrote:
> >  > 
> >  > > Hi RDF Interest,
> >  > > 
> >  > > Feel free to shoot me down because I'm way out of my depth here, but I
> >  > > was wondering:
> >  > > 
> >  > > Would it be possible to build a tool that could take an owl full
> >  > > ontology, infer a set of owl DL compatible assertions and then
> >  > > remove all the non-OWL DL compatible ones?
> >  > 
> >  > Yes. But you need a decent set of OWL DL tools to motivate the amount of 
> >  > work it would be to build such a thing.
> > 
> > True, but it seems that despite there being a lack of DL tools, people
> > still want to restrict themselves to OWL Lite/DL.  
> > 
> > What's your view on tjis subject? In the absence of DL tools, should
> > people be targetting OWL full by default?
> > 
> > Cheers,
> > 
> > Phil
> > 
> > 
> 
> My view is that I planned to do the work that you described, but could
> not justify the three or four months of my time that I thought it would
> take without a DL toolset.
> 
> i.e. to build OWL DL rather than OWL Full costs more money, because you
> have to ensure that you are in the DL subset, without enough DL tools to
> justify this expenditure it is a waste of money and effort that could
> better be spent elsewhere.
> 
> In as much as my decision to not do such work was made in collaboration
> with my colleagues, manager and Jena users, this view represents a wider
> opinion than just my own.
> 
> I, personally, also find the OWL DL story fairly compelling, but without
> some substance behind it, from the DL community, cannot actually support
> it with my own effort.
> 
> Jeremy

Anyway,  Phil Dawes's question is worthy of noticing. My thought:

For a given owl ontology (may be in OWL full), it's useful to extract the OWL Lite and OWL DL subsets of the given ontology, and then (automatically) recommend an OWL Lite ontology (as well as an OWL DL ontology)  that is close to (very similar to) the original one.

The first step is mainly at syntax level (as Jeremy Carrol pointed out before), while the second step is at semantics level and may use some matching techniques.


Yuzhong Qu

Received on Monday, 19 July 2004 22:49:46 UTC