W3C home > Mailing lists > Public > public-webont-comments@w3.org > March 2002

OWL and Published Subjects - is coordination relevant?

From: Bernard Vatant <bernard.vatant@mondeca.com>
Date: Tue, 26 Mar 2002 19:58:38 +0100
Message-ID: <005a01c1d4f8$3ecbb500$ea91fea9@bernard>
To: "tm-pubsubj-comments" <tm-pubsubj-comment@lists.oasis-open.org>, <public-webont-comments@w3.org>
I posted last week a call for reflection to both PubSubj an WebOnt working groups
http://lists.w3.org/Archives/Public/public-webont-comments/2002Mar/0011.html

I would like to expand a little more on why I see some liaison between OASIS PubSubj and
OWL would be relevant from both viewpoints. For example several points in the OWL
Requirements document focus on the necessity of interoperability of ontologies.

<< 3.1 Shared ontologies:
Interoperability requires agreements on the definitions of terms. Ontologies can provide
standard sets of terms and formal
descriptions of those terms. Data sources that commit to the same ontology explicitly
agree to use the same terms with the
same meanings. >>

And also

<< 3.3 Ontology interoperability
Different ontologies may model the same concepts in different ways. The language should
provide primitives for relating different
representations, thus allowing data to be converted to different ontologies and enabling a
"web of ontologies."

Although shared ontologies and ontology extension allow a certain degree of
interoperability between different organizations and
domains, there are often cases where there are multiple ways to model the same information
... >>

How can those agreements and differences be clearly expressed? Referring to a definite
term in a non-ambiguous way needs this term to be identified independently both of its
name, and of the relationships declared in any given ontology it happens to belong. PSI
are fit to provide such ways for identification, clarification and hence non-ambiguous
agreement.

To refer to the example of Dolphins quoted in the Requirements doc, a PSI could be defined
e.g by http://psi.seaworld.org/animals/dolphin.html
This hypothetical PSI should provide a non ambiguous definition of what Dolphins are ...

One zoological ontology could use this PSI to identify the "Dolphins" class in the generic
class hierarchy
Animals > Mammals > Cetaceans > Dolphins

Note that the French version will use the same PSI to identify "Dauphins"
Animaux > Mammifères > Cétacés > Dauphins

Another ontology used by a French Riviera tourism website could use the same PSI to
identify the "Dauphins" class in a more local and specific hierarchy
Antibes > Attractions > Marineland > Dolphins

Referring to the same PSI for those two classes will support assertions like:

"If Flipper is an instance of Dolphins in Marineland, then he is a Mammal"
"Instances of Cetaceans can be seen in Marineland, Antibes"

But of course, I would make a mistake using the same PSI to identify my local swimming
team :))
Embrun > Sports > Natation > Clubs > Les Dauphins

BTW, it figures that URLs of Open Directory categories - a use case quoted in the
Requirements doc - are not fit to be used as conformant PSIs, since they are defined in
their very syntax by the category hierarchy, see for example the following.

http://dmoz.org/Science/Biology/Flora_and_Fauna/Animalia/Chordata/Mammalia/Cetacea/Toothed
_Whales/Dolphins/
http://dmoz.org/Recreation/Travel/Specialty/Ecotourism/Dolphins,_Manatees,_and_Whales/Dolp
hin_Swimming/
http://dmoz.org/Kids_and_Teens/School_Time/Science/Living_Things/Animals/Mammals/Dolphins_
and_Porpoises/

Go figure if Dolphins in three sub-sub ... categories are the same guys, even if those
three categories are cross-referencing each other. Not only the URIs are stuck to a
specific hierarchy, but moreover, they are notoriously unstable (I've been there -
redefining hierarchies is a popular fun game among Open Directory editors)

To come back to our working groups, I think it would be great to explore at least two
tracks:

1 - How OWL could integrate the notion of PSIs?

2 - How PubSubj Recommendations could take this very issue of ontologies interoperability
as an explicit use case?

And since our two efforts are roughly at the same stage of development (more issues than
definitive answers), maybe they could be able to achieve recommendations consistent with
each other, or at least explore together their common issues.

Thanks to consider (from both sides)

Bernard

Bernard Vatant
Chair - OASIS TM PubSubj Technical Committee
www.oasis-open.org/committees/tm-pubsubj/
Received on Tuesday, 26 March 2002 14:00:35 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:43:27 GMT