Re: OWL-S version 1.1 now available

On Nov 20, 2004, at 7:31 PM, Ian Dickinson wrote:

> Hi Bijan,
>> Thanks for the read and comments!
> You're welcome.
>
>>>  The namespace is .../generic/ObjectList.owl, so perhaps objList: or 
>>> generic: would be more graceful namespace prefixes?
>> Er..generic...definitly not :) objList perhaps, but I also intended 
>> it to read fairly closely with the ormal rdf prefix, e.g., rdf:List, 
>> shadow-rdf:List (i.e., it's a sort of rdf:List-a-like).
> I guess my comment was founded on the fact that *everyone* who looks 
> at rdf:List and wants one in their ontology has the same problem.

Yes? And we've provided a nicely generic solution.

>  It's certainly something that comes up fairly frequently on the Jena 
> forum. So owl-s is not alone in wanting to "shadow" rdf lists, and 
> therefore it seemed to me that it would be better to be a little more 
> self-contained.

Er...I'm finding a lot of a non in this sequitur. The URI for the 
namespace was out of my control, but I don't much care. It's perfectly 
generic. The "canonical" prefix was picked (by me) to make it clear 
what was going on (i.e., we were shadowing the rdf list vocabulary). If 
you also want to shadow the collection vocabulary, please feel free to 
use this lil ontology.

What it's *not* doing is providing a semantically correct 
representation of lists, a la Lisp lists. It's my understanding that 
you can't really do that in first order logic (since you need well 
foundedness and transitive closure). There are some other issues (e.g., 
whether it's harmful to not be able to rule out the deviant lists).

[snip]
>> But, I think it's as graceful, acutally rather more so, than your 
>> proposed alternative (*especially* generic).
> :-)  I usually try to make my namespace prefixes from the, er, 
> namespace they represent.

I don't feel so constrainted.

> Since generic was a part of the namespace it seemed worth a throwaway 
> suggestion.  *I* wouldn't use generic as a namespace prefix in my RDF 
> documents either, as it happens!

Heh.

I still like shadow-rdf. you have rdf:List and shadow-rdf:List. Verra 
clear which is which and how they're intended to be releted.

FWIW, the shadow-list approach was intended as a fast stop gap before 
reworking how components were to be represented. I prefer using a 
hasComponent like property, with ordering to be determined by some  
property of the performs.

Cheers,
Bijan Parsia.

Received on Sunday, 21 November 2004 10:08:46 UTC