RE: OWL-S: Parameter modeling

Bijan,

Thanks for your detailed response. I understand now why I am in OWL-Full. Basically the approach I suggest, is using second-order predicate (meta-property). Note that Cyc is also using working second-order logic reasoner using second-order predicate. Ref. http://www.cyc.com/doc/tut/ppoint/SecondOrderPredicates_files/frame.htm
What is wrong with their approach ? 

I think when it comes to model function or process, adding the support of second-order predicate in reasoners will simplify considerably the description of process or functor. The question is whether this addition of second-order predicate (a small subset of second-order predicate logic) would keep the reasoner decidable and can compute in a finite time.  
 
Why should we stick religiously to DL ?  Not being an implementer of inference engine, I am asking whether it is a huge task to add this second order predicate extension in existing DL reasoner or not ? 


[snip]
>(Bijan) So this is about execution traces?

Yes it is. How does OWL-S solve this problem today ?

Best regards
 
Stephane Fellah
Web Chief Architect
 
PCI Geomatics
490, Boulevard St Joseph
Hull, Quebec
Canada J8Y 3Y7
Tel: 1 819 770 0022 Ext. 223
Fax 1 819 770 0098
Visit our web site:  www.pcigeomatics.com
 


-----Original Message-----
From: Bijan Parsia [mailto:bparsia@isr.umd.edu] 
Sent: Friday, July 02, 2004 11:52 AM
To: Stephane Fellah
Cc: public-sws-ig@w3.org
Subject: Re: OWL-S: Parameter modeling

I had trouble with your formatting. I prefer non-HTML/rich text email 
just for such reasons :)
[snip]
> (Stephane) Please note that I am not an expert in DL, so please point 
> me out where my reasoning is wrong.
>
>  rdf:Property is the class of RDF properties. rdf:Property is an 
> instance of rdfs:Class.

Not in OWL-DL.

> So theorically, owls:parameter, a subclass of rdf:Property can have 
> properties specific to Parameter (name,description, binding...)

I recommend the various documents from the OWL working group.

[snip]

> (Stephane) I consider parameter as a property with a specific role.

Great. I don't. Necessarily.

>  Thus I assume it is a subclass of RDF Property with additional 
> properties.

I don't believe that's necessarily the best modeling choice. However, 
if that's your style, you're pretty much committed to OWL Full,

> The most important property of a Parameter is its range (which is 
> called parameterType in owl-s 1.x). For me parameterType = range or is 
> a subproperty or range.  

Yes, this way of reifying things works that way.
[snip]
> (Stephane) Property is a Class also, thus can be extended by 
> properties.

As I said, not in DL.
[snip]
> (Stephane) Why the last two properties would have to be 
> AnnotationProperties ? Does the definition below not valid in OWL-DL ? 
> I declare the metadata property as a datatypeProperty, not an 
> annotationProperty.

Because in OWL DL meta properties (properties of properties or of 
classes) have to be annotation properties. There is a more elaborate 
justification, and there are ways of relaxing some of that (annoation 
properties are, themselves, a relaxation of standard first order 
representations).

>  You are right when you say that my last two properties are 
> metaproperties. Does this mean I am in second-order predicate logic ? 
> (Once again I am not expert). Why current validator would not work in 
> this case ?

I gave some references in my last note. Check them out.

You might look at:
	http://www2003.org/cdrom/papers/refereed/p050/p50-horrocks.html
[snip]
> (Stephane) This represent an instance of a process with the parameter 
> bound to some value. ParameterValue does not need to be model as a 
> class. It the Object/Literal part of the triple. This notation is very 
> useful to perform the lineage of a complex process (think about a 
> complex image processing chaining task). The operator wants to keep 
> track of the parameters applied at each step. Using my approach this 
> would have a compact syntax.\

So this is about execution traces?

[snip]

> (Stephane) These documents are too theorical for me, and I may not 
> understand all the details of DL theory. Do you have a short answer 
> for this. Is it second order predicate logic or not ? Is it OWL-Full 
> or not ?

It's very hard to debate sensibly if you're unwilling to acquire the 
relevant background.

It is OWL-Full.

It is not (necessarily) second order. You can reify all the relations, 
a la the DAML+OIL axiomatic semantics. E.g., isntead of s p o being 
mapped to P(s,o), you map it to triple(s, p, o).

[snip]
> (Stephane) I think both approaches are valid. However I think the 
> first approach (pre-1.0) modeling parameter as property is more 
> natural and readable than the second one (Parameter as class).  

I don't think any of this is particular natural, myself.

Readable, perhaps, but I don't think that's an interesting criterion 
since I don't think any OWL modeling of a syntax tree is going to be 
usefully readable. I put my faith in a surface syntax.

Cheers,
Bijan Parsia.

Received on Friday, 2 July 2004 14:08:28 UTC