W3C home > Mailing lists > Public > www-rdf-logic@w3.org > March 2003

Re: Even more Fuzzy about FunctionalProperty!

From: Roger L. Costello <costello@mitre.org>
Date: Wed, 05 Mar 2003 08:33:28 -0500
Message-ID: <3E65FCA7.B9AD27AA@mitre.org>
To: www-rdf-logic@w3.org
CC: "Costello,Roger L." <costello@mitre.org>

Thanks Mike.  I am not real comfortable that you have completely
answered my question, so let me followup with the example that you
presented.

You have stated that "hasFather" is a FunctionalProperty.  Consider
these two instances:

<Person rdf:ID="Mike">
    <hasFather>
        <Person rdf:ID="Joe">
            <age>56</age>
        </Person>
    </hasFather>
</Person>

<Person rdf:ID="Mike">
    <hasFather>
        <Person rdf:ID="Joseph">
            <address>101 Curl Drive, Columbus, OH</address>
        </Person>
    </hasFather>
</Person>

Since "hasFather" has been declared to be a FunctionalProperty can we
infer that this:

        <Person rdf:ID="Joe">
            <age>56</age>
        </Person>

is equivalent to this:

        <Person rdf:ID="Joseph">
            <address>101 Curl Drive, Columbus, OH</address>
        </Person>

Thus, when inferencing do we treat the values of "hasFather" as a "black
box", or must we "reconcile" the values?  /Roger


Mike Dean wrote:
> 
> A better example for FunctionalProperty is probably
> hasFather.  If I know that
> 
>   :hasFather rdf:type owl:FunctionalProperty
>   :mike :hasFather :joe
>   :mike :hasFather :joseph
> 
> I can then conclude that
> 
>   :joe owl:sameIndividualAs :joseph
> 
> Such cardinality reasoning with data type properties is less
> useful, although the modelling aspects of "at most 1" can
> also be very important (as seen in most programming
> languages and database systems).
> 
> Also, I would recommend that folks avoid using a value like
> "6300 kilometers" in favor of a representation that made
> 6300 an XML Schema numeric datatype (e.g. xsd:decimal)
> suitable for arithematic computation and :kilometers an
> instance that could then be related to other instances such
> as :miles and :meters.
> 
>         Mike
Received on Wednesday, 5 March 2003 08:32:28 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:52:43 GMT