Re: [OWL-S] DLization of 1.1

On Apr 7, 2004, at 9:07 AM, Daniel Elenius wrote:

>> My current thought is to add some decoration to RDF/XML that would 
>> let an XSLT script know which parseType=Collection to covert. Since 
>> we'll want something like that for data value oneOfs anyway, maybe 
>> it'll get some traction.
>>
>> The ugly part is you need to do a bunch a work to hand data valued 
>> lists.
> Let me see if I get this right. Are you suggesting that OWL-S files 
> are first written using something similar to parseType=Collection plus 
> some extra decoration, and then converted using XSLT to a 
> representation that is proper OWL DL?

Yes.

>> Yep. A good trick. One we discussed. There was no consensus for 
>> anything new this round. I hope we make such a decision for the final 
>> version.
>>
> Cool.

Well, either an indexed list choice, confirm the shadow list, or use 
some other representation like before/after. Suggestions welcome.
[snip]
> An alternative would be to use "singleton" class instances, like so:
>
> <owl:Class rdf:ID="SignInData">
>  <owl:oneOf rdf:parseType="Collection">
>    <owl:Thing rdf:about="#TheSignInData"/>
>  </owl:oneOf>
> </owl:Class>
>
>
> <process:Input rdf:ID="CongoBuySignInInfo">
> <process:parameterType rdf:resource="#TheSignInData"/>
> </process:Input>
>
>
> But I'm not sure if these inputs are intended to be instantiated or 
> not. If they are, then this doesn't work.

They are, so it doesn't work.

>  But it looks to me like we just want to point to a data _type_ rather 
> than an instance _value_, in which case my suggestion would work, I 
> think.

We do want to point to a datatype *but* we want (well, some of us want 
;)) people to be able to point to ordinary OWL classes.

>  This way of doing it means more semantic information in the model 
> without a dedicated reasoner to extract the datatype given as a URI, 
> as you designed it.

Not really. What more information? Actually, I completely miss the 
advantage. It's like declaring a new type of type, owl-sType, and 
having instances rdf:value particular classes. Worse that that, 
actually, since that's a pretty good solution now that I think about it 
:) It fits in with how we're ding a *lot* of things.

> Unfortunately though, with my way of doing it, different kinds of 
> parameterType are needed for datatypes and OWL classes, since it can't 
> be both a DatatypeProperty and an ObjectProperty.

Yep. But with a wrapper object, that wouldn't be a problem.

This definitely requires more thought. One nice thing about the current 
solution is that it's not *that* far off the older one, is reasonably 
clean, and wouldn't be that hard to change later.

That said, good solutions welcome for 1.2.

Cheers,
Bijan Parsia.

Received on Wednesday, 7 April 2004 09:16:41 UTC