Re: Examples of Syntax and semantics

With regard to this solution.
I understand that you have mapped one set of statements into the other using
the pre-existing vocabulary of the given ontologies and their comon parents,
wich I assume has proved the time dependant syntax?
A couple of things occur to me.
Would it be possible for a machine to do this? I know that once worked out,
a machine can utilize it, but that is not the same. What I am asking is how
would a machine resolve arbitrary ambiguity?
I would have thought that this is where a third, well thought out, ontology
would come in.
What is the point, in a case such as this, of not devising such an ontology?
You solution seems relatively complex for a person to implement (compared to
other operations in using, say foaf, or, even, extending foaf).
Although this example is straight forward enough to follow, others would not
be and this is also where a parent ontology would have a role.
As Tim said:-

an "Upper Level" ontology which introduces new concepts (divorce,  death,
> ceremonies etc)
>

which is more intuitive to a human creator and user. This is important, not
just because it makes the schema easier to grok, but because it can lead to
further developments, that is uses, elaborations and enhancements.
So, it seems to me, there is a dynamic tension between the "let's get on and
code" and the "let's model" lot. This is a bit strange as coding and
modeling are intertwined in working with these languages.
If there is a team, it would probably need both sorts and to live with the
tension between them.
But if the argument is being made about a point of principal ... ?
Adam
On 07/04/06, Henry Story <henry.story@bblfish.net> wrote:
>
>
> On 7 Apr 2006, at 11:04, <tim.glover@bt.com> <tim.glover@bt.com> wrote:
> > Henry,
> >
> >>> Henry,
> >>> Thank you for the useful reply.
> >>> But I don't think this deals with the issues of mediation between
> >>> two similar but semantically disjoint ontologies.
> >
> >> Show me one, and I'll look at how we can link them (if it does not
> >> take too long)
> >
> >
> > OK, here is one attempt at a challenge (adam - sorry if I have missed
> > the point)
> >
> > ********************************************
> > Imagine 4 ontologies. All describe the class People, and subclasses
> > Married and Unmarried.
> >
> > Suppose that Clive was married but is now divorced. Diane was married
> > but is now widowed. Alice has never been married, and Bill is married.
>
>
> I have added below the statements in N3. The reason I add the graphs
> is just to clarify where the data was fetched from (after all we
> don't just
> get statements out of the blue). This is especially important as we have
> ontologies that are time dependent.
>
> > * In ontology 1, Clive and Diane are unmarried (unmarried means
> > unmarried now)
>
> {
> ppl:Clive a o1:Unmarried .
> ppl:Diane a o1:Unmarried .
> } :retrieved [ :at "2006-03-23T10:00:00Z";
>                 :from <http://eg.org/couples1> ] ;
>    :name :g1 .
>
> > * In ontology 2, Clive and Diane are married (married means has gone
> > through a marriage ceremony).
>
> {
> ppl:Clive a o2:Married .
> ppl:Diane a o2:Married .
> } :retrieved [ :at "2006-03-23T10:00:00Z";
>                 :from <http://eg.org/couples2> ] ;
>    :name :g2 .
>
> > * In ontology 3, Clive and Diane are both married and unmarried
> > (married
> > means has been married once, unmarried means not married now)
>
> {
> ppl:Clive a o3:Married, o3:Unmarried .
> ppl:Dian  a o3:Married, o3:Unmarried .
> } :retrieved [ :at "2006-03-23T10:00:00Z";
>                 :from <http://eg.org/couples3> ] ;
>    :name :g3 .
>
> > * In ontology 4 Clive is unmarried, Bill is married, and Alice and
> > Diane
> > are neither married nor unmarried (un-married means having been
> > through
> > a divorce. Married means married now)
>
> {
> ppl:Clive a o4:Unmarried .
> ppl:Bill a o4:Married .
> ppl:Alice a o4:NeitherMarriedNorUnmarried .
> ppl:Diane a o4:NeitherMarriedNorUnmarried .
> } :retrieved [ :at "2006-03-23T10:00:00Z";
>                 :from <http://eg.org/couples3> ] ;
>    :name :g4 .
>
> >
> > How do you map these ontologies?
> >
>
>
> If someone has been married at a time, then they have been Married once.
>
> # Both o1 and o3 are equivalent. They mean unmarried now.
> o1:Unmarried owl:equivalentClass o3:Unmarried .
>
> # If you have gone through a Marriage ceremony then you have been
> married I suppose
> # otherwise ignore
> o3:Married owl:equivalentClass o2:Married .
>
> #you can't be married now and unmarried now
> o4:Married owl:complementOf o1:Unmarried .
>
> # someone who has been through a divorce was married once
> { ?p a o4:Unmarried . } => { ?p a o3:Married . } .
>
> # let us defined o4:NeitherMarriedNorUnmarried
>
> o4:NeitherMarriedNorUnmarried owl:complementOf [ owl:unionOf
> (o4:Married o4:Unmarried ) ] .
>
> # Some of the classes are indexicals. Let us mark the time dependend
> classes as follows:
>
> o1:Unmarried owl:subClassOf tm:TimeDependent .
> o3:Unmarried owl:subClassOf tm:TimeDependent .
> o4:Married owl:subClassOf tm:TimeDependent .
>
> Because we have fetched all the above graphs simultaneously, we can
> merge them together
> and place them all in one big time dependent graph.
>
> I am sure one could do more. But perhaps there are some tests you
> would like us
> to try out?
>
>
>
>
> > ***********************************************
> >
> >> That notion is one that we keep telling you is not part of the
> >> Semantic Web. It is *your* vision of the Semantic Web.
> >> You have a straw man argument. You imagine we are doing something we
> >> are not doing. Then you prove that what you imagine we are doing is
> >> impossible. And you wrongly conclude that what we are doing is
> >> impossible. Please distinguish between what you think we are doing
> >> and
> >
> >> what we are doing.
> >
> > Who are "We" exactly? People that you happen to agree with? ;-)
> >
> > Tim.
>
>
>
>
>