Re: Why has the RDF mapping conditional mapping rules? (ISSUE-68)

On 6 Feb 2008, at 16:42, Michael Schneider wrote:

> Hi again!
>
> I hope that I did not produce more cofusion with my yesterday  
> night's mail
> then without posting it. :) There might simply be a  
> misunderstanding on my
> side, since I did not deal much with this question until now.
>
> Perhaps, did I misunderstand the intention behind roundtripping in  
> general,
> and the backwards compatibility issue of the RDF mapping in special?

I think there's an interaction between roundtripping and backwards  
compatibility, although they are independent concepts.

Remember that the functional syntax specs an API (see the OWL API for  
a concrete implementation). So, it would be unfortunate if the  
"default" FS2RDF mapping resulted, of necessity, in something that  
was *not* OWL 1.0. If that were the case, we'd be giving very bad  
advice to implementors as going from OWL/RDF1.0 to FS to OWL/RDF  
again would result in something fairly unrecognizable. I don't mind  
offering a switch to "upgrade" to a more typed style, but we cannot  
force that by default (IMHO).

> Was it
> the intention to get an OWL-1.0-DL ontology, whenever the  
> functional syntax
> in 1.1 "corresponds strongly" to some abstract syntax in 1.0? So  
> that new
> RDF vocabulary is used only when either a new 1.1 language  
> construct is
> used, or when punning is used?

Yes.

> I think this would approximately amount to a
> roundtripping the other way around than I thought, from 1.0/RDF to  
> 1.1/FS
> and back to 1.0/RDF. Was this the idea?

We certainly want to support that. After all, most owl ontologies  
existing now are OWL 1.0. And most tools expect OWL 1.0. It doesn't  
make sense to make a change when there's no gain. (Changing when  
forced to for technical reasons is still *unfortunate* but at least  
justifiable.)

Cheers,
Bijan.

Received on Wednesday, 6 February 2008 16:56:00 UTC