Re: Where are the semantics in the semantic Web?

No system can be entirely self-describing, so what ultimately matters 
are the set of axioms and the logic. When dealing with real world 
systems the set of axioms is extremely large - common understandings. In 
practical terms then, there is a logic that can be applied to a set of 
assertions that are grounded in external agreements. Semantic 
interoperability is a function of how these agreements are reached and 
recorded and understood.

There is a "position paper" of a team looking at semantic 
interoperability issues in the Marine sector, at:

http://www.iode.org/marinexml/files/MarineXML_Position_Paper_R4r0.pdf

The conclusion is that the usefulness of any semantics ultimately 
derives from the governance process behind the assertions. Pragmatic 
epistemology seems to have utility: to scale the solution we need 
mechanisms for independence, but also the mechanisms to merge or map 
between ontologies to avoid unecessary duplication of effort and 
creation of artificial ambiguities.

Our next path to pursue is the feasibility of interoperation of the 
semantic registries across governance domains. This would seem to 
provide the machine-readable semantics required to exploit 
relationships, whilst allowing the richness and variability of 
alternative viewpoints for different semantics of use. My own feeling is 
that a rich enough soup of practical definitions built from the bottom 
up, with a top-down coherence through commonality of mechanism, would 
eventually provide opportunities for stochastic or heuristic natural 
language oriented approaches.

Regards
Rob Atkinson
Social Change Online

Shi, Xuan wrote:
> Here is another good article by Michael Uschold published by the AI
> Magazine. Also you may wish to read it at:
>
> http://www.findarticles.com/p/articles/mi_m2483/is_3_24/ai_110575581/print
>
> Please pay attention to the "law of the semantic web" as Uschold concluded:
>
> "The more agreement there is, the less it is necessary to have
> machine-processible semantics" 
>
> I am impressed by such statement. This means to me as that we need more
> standards and agreements rather than logics to develop semantic Web
> services. People use logics because one cannot understand each other, then
> we have to guess what the content means by using logics to match the content
> of the service. However, what and how machine can process depends on what
> and how human beings design it. Computer itself does not have powerful
> logical ability to understand varied kinds of semantics in the Web services
> in different domains with complex business models.
>
> However, if service semantics is developed based on standards and
> agreements, then everything is clear and we do not need logic for
> matchmaking.
>
> For example, in the airline ticket reservation service, since all of us know
> the requirements of the input varibles and prospective outcomes, then such
> common sense can be the foundation for domain specific standards. This
> means, if service requesters send a request with required and detailed
> information about departure/arrival city or airport, time schedule, etc. to
> any service provider, they should get the response in the same format. 
>
> For each single company, such as US Airline, United, etc. such response only
> have the ticket information of its own service. However, for service
> providers like expedia, requesters can send out the same request and get a
> response that contains ticket information of multiple companies in the same
> format/template.
>
> If service semantics can be formulated based on standards, then service
> requesters may wish to send multiple functional requests in one single XML
> document to service providers like expedia to reserve tickets, hotels,
> rental cars, etc. and will get back the response also in one single XML
> document. Service requesters do not need to know and care about how provider
> process their request but can get a response to match their request.
>
> If there is no standard or agreement in such process, we have to guess the
> meaning of the service by using different logics.
>
> At last, here is the link to an article "Interoperating GISs" published in
> 1997 for your kind attention:
> http://www.ncgia.ucsb.edu/conf/interop97/report.html
>
> Regarding technical, semantic, and institutional problems in GIS
> interoperability, it is said in this document that "The semantic level of
> interoperability addresses these issues of shared meaning, and is clearly
> more problematic than the technical level". When Web service (in a broad
> viewpoint) technology can solve the technical problems, we now face the
> semantic problems. The experience in GIS community is just to develop
> industrial standards and this is the main task and interest of OGC (Open GIS
> Consortium). But now, we face the problem to sharing GIS knowledge and
> methodologies beyond GIS community via semantic Web services and get more
> problems.
>
>
>
> -----Original Message-----
> From: Cutler, Roger (RogerCutler)
> To: Shi, Xuan; Dan Brickley ; drew.mcdermott@yale.edu
> Cc: public-sws-ig@w3.org
> Sent: 11/21/05 3:24 PM
> Subject: RE: Options we have with respect to the draft charters (i.e.,  RE:
> [fwd] Draft charters for work on Semantics for WS)
>
>
> Very interesting paper.  Thank you. 
>
> -----Original Message-----
> From: public-sws-ig-request@w3.org [mailto:public-sws-ig-request@w3.org]
> On Behalf Of Shi, Xuan
> Sent: Monday, November 21, 2005 10:05 AM
> To: 'Dan Brickley '; 'drew.mcdermott@yale.edu '
> Cc: 'public-sws-ig@w3.org '
> Subject: RE: Options we have with respect to the draft charters (i.e.,
> RE: [fwd] Draft charters for work on Semantics for WS)
>
>
> I recommend this 87-page paper "Towards a Semantic Web for Culture" by
> Kim H. Veltman which can be accessed at:
>
> http://jodi.tamu.edu/Articles/v04/i04/Veltman/veltman.pdf  
>
> The so-called "Semantic Web" in nature is "logical Web", the result is
> even XML people cannot understand RDF/OWL due to those logics and the
> way of RDF presentation. That's why this technology is not well accepted
> and deployed.
> That's why I said here before, the more complex the system, the less the
> user. It's the same to developing semantic Web services.
>
>
>

Received on Tuesday, 22 November 2005 05:15:26 UTC