Re: Inheritance and RIF

Hans--

I'd meant to ask this before:  what exactly do you mean by 
"inheritance"?  As Peter noted in his reply, what OWL supports is the 
subclass relationship, so that if X is defined as a subclass of Y, x's 
are defined to be y's, and certain inferences about the x's can be made 
as a result. If you mean the object-oriented programming "inheritance" 
(implementation-reuse, sometimes with overriding), OWL doesn't support that.

--Frank

Hans Teijgeler wrote:
> Peter and Frank,
> 
> This forum seems to recover from the shock of the raid by the ONTACians. 
> 
> I suggest that we return to more mundain subjects, like inheritance:
> 
> If inheritance is taken care of at all in OWL and/or OWL-related software,
> where then exactly is it being taken care of? Is it:
> - in the language, hidden in the S&AS?
> - in "standard" reasoners or other software?
> 
> If it is defined in the S&AS or similar, I guess that in order to claim
> compliance one must adhere to that.
> 
> Or is everybody on its own (which is fine with me)? (Yes Frank, we do use
> n-ary relations all over the place, so we can handle it. I haven't heard
> from Evan Wallace.)
> 
> Regards,
> Hans
> 
> -----Original Message-----
> From: Frank Manola [mailto:fmanola@acm.org] 
> Sent: Monday, April 03, 2006 18:24
> To: Hans Teijgeler
> Cc: semantic-web@w3.org
> Subject: Re: Inheritance and RIF
> 
> Hans Teijgeler wrote:
>> Peter,
>>
>> Thanks for your response!
>>  
>> You wrote:
>>
>> <PFPS>
>> Well, RIF shouldn't have any mechanism for inheritance.   After all, RIF
> is
>> not about classes and instances, but is instead about rules.
>> </PFPS>
>>
>> I hope that those rules are about anything we know of, i.e. about 
>> classes, individuals, and properties. What else can a rule be about?
>>
>> I still am in need for an answer to my question: does anyone know how 
>> to represent a value with a tolerance in RDF/OWL? So something like: 
>> diameter = 150 mm +0.03%  -0.25%.
> 
> http://www.w3.org/TR/swbp-n-aryRelations/ discusses a use case for
> describing "Christine has breast tumor with high probability".  The same
> approach (basically, n-ary relations) could be used to describe values with
> upper and lower tolerances as well.  This may not be the way manufacturing
> folks would do it, but it's certainly possible.  NIST does a lot with
> manufacturing (and Evan Wallace is on the SW Best Practices group);  perhaps
> someone from there could comment on this.
> 
> --Frank
> 
>> XML Schema doesn't have any datatype that can do that. I assume that 
>> somewhere in SemWebland someone must have been dealing with tolerances 
>> and accuracies. After all it is necessary when "making one of the 
>> standard modelling choices", i.e. that of defining classes by means of 
>> "criteria for membership". Are there other modelling choices?
>>
>> Kind regards,
>> Hans
>>
>

Received on Tuesday, 11 April 2006 21:34:39 UTC