W3C home > Mailing lists > Public > www-ws@w3.org > October 2003

[owl-s] Re: Data Flow

From: Bijan Parsia <bparsia@isr.umd.edu>
Date: Thu, 9 Oct 2003 22:19:45 -0400
To: www-ws@w3.org
Message-Id: <37374280-FAC8-11D7-82AF-0003936A0B26@isr.umd.edu>

On Thursday, October 9, 2003, at 09:59 PM, Massimo Paolucci wrote:

> Bijan Parsia wrote:
>
>> On Thursday, October 9, 2003, at 08:50 PM, Massimo Paolucci wrote:
[snip]
>>> A somewhat better way to do the same thing is to define a new 
>>> property
>>> "dataLink" that is a subproperty of sameAs but is restricted to 
>>> relate
>>> outputs to inputs.
>>
>>
>> I think this is illegal in DL. FWIW. In some sense, it might be "ok" 
>> in the sense that it's really just a syntactic trick. But I would 
>> want to check. I'm less adverse to doing things that are "DLable" but 
>> happen not to be DL, than I am to doing robustly OWL full things.
>
> Actually,  the way I understand sameAs, it does not push the ontology 
> directly in OWL Full, unless it is used to equate classes.  The OWL 
> Reference states...
>
>            In OWL Full, where a class can be treated as instances of 
> (meta)classes, we can use
>            the |owl:sameAs| construct to define class equality, thus 
> indicating that two concepts
>            have the same intensional meaning.
>
> This is not what we are doing here.

That's not why you'd be in owl full. It's because of redefining the 
builtin vocabulary (or subClassing or subPropertying it).

If you look at the transformation table:
	http://www.w3.org/TR/owl-semantics/mapping.html#4.1
You'll see that sameAs doesn't get mapped to a property in the abstract 
syntax but to a syntactic construct. Ergo it can't be part of a 
subPropertyOf construct.

Ergo, not in OWL DL. BTW, OWL reference isn't normative and I don't 
know if there's a single place where everything that would put you in 
owl full is clearly listed.

I recomment using Sean's validator or pellet or both.

Oh a better citation:
	http://www.w3.org/TR/owl-semantics/mapping.html#disallowed_vocabulary

""" When considering OWL Lite and DL ontologies in RDF graph form, care 
must be taken to prevent the use of certain vocabulary as OWL classes, 
properties, or individuals. If this is not done the built-in 
definitions or use of this vocabulary (in the RDF or OWL specification) 
would augment the information in the OWL ontology.
...
The disallowed vocabulary from OWL is... owl:sameAs..."""

Q.E.D :)

Your *use* of sameAs with individuals is fine. Your *subpropertying* of 
sameAs is not.

Which, uh, is what I said before :)

>> OTOH, if we have a bonafide surface syntax that diverges plenty from 
>> the owl, that's the right place to put the macro.
>>
>>> This property is defined as follows.
>>
>>
>>> <owl:ObjectProperty rdf:ID="dataLink">
>>>  <rdfs:subPropertyOf rdf:resource="&owl;#sameAs"/>
>>>  <rdfs:domain rdf:resource="#Output"/>
>>>  <rdfs:range rdf:resource="#Input"/>
>>> </owl:ObjectProperty>
>>>
>>>
>>> The advantage is that we introduce some sort of directionality to 
>>> data
>>> links,
>>
>>
>> Except, uh, 1) sameAs is symmetric, right? Plus, anything on either 
>> side will be *both* an Input and an Output. So no directionality. 
>> Nothing *but* identity. You can't be a little bit sameAs.
>
> First of all sameAs is a property and it is not defined as symmetric 
> in the OWL specs

If you look at the semantics, you'll see, I think, that it is 
symmetric. It may not be an owl:SymmetricProperty, but, then again, in 
OWL DL, it isn't defined by an owl ontology in the first place.

>  (although I agree that that would make a lot of sense).  But in 
> general the attempt is to introduce constraints so that we can detect 
> meaningless patterns.

I understand. But it fails to do so in OWL DL. This may justify, in 
your and the rest of the coalition minds moving to OWL Full, but it 
doesn't stay in OWL DL.

[snip]
>> To sum, I think the subpropertying is non DL (FWIW) and the 
>> subPropertying doens't have the effect you want. However, the basic 
>> proposal may work, and well! But I'm not entirely sure and would wan 
>> to develop some examples.
>
> I will try to send an example tomorrow.

Great!

>> I think it doesn't conflict with my parameterType proposal, either.
>
> I sort of like this ;-)

Even better :)

Cheers,
Bijan.
Received on Thursday, 9 October 2003 22:21:07 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 3 July 2007 12:25:44 GMT