W3C home > Mailing lists > Public > public-semweb-lifesci@w3.org > April 2006

RE: Ontology editor + why RDF?

From: <wangxiao@musc.edu>
Date: Mon, 3 Apr 2006 10:01:17 -0400
Message-ID: <20060403100117.o4fodiyy0og04k8s@webmail.musc.edu>
To: public-semweb-lifesci@w3.org

Quoting "Kashyap, Vipul" <VKASHYAP1@PARTNERS.ORG>:

>
>
> A straight-foward "porting" is always wrong, or lack of the comprehension of
> difference between RDF and other exiting technologies.  RDB and XML schema
> contains implicit semantics that should be explicitly expressed in RDF.
>
>
>
> [VK] Making the underlying data model constraints in the RDF graphs is what I
> meant by extra semantcis - In the relational model, that would correspond to
> primary, foreign keys, functional and inclusion dependencies, etc.... Some of
> this would likely require OWL...

I think your definition of semantics differs from mime.  I do not think that
only terms defined in OWL has semantics. When you make a statement, or in your
words, cross-linkings, you express your semantics.  Such cross linking is
different than an HTML page is "cross-linked" to other pages.

When I say
(1) _:foo ex:differs _:bar .

I can express the same semantics as

(2) _:foo owl:differentFrom _:bar .

You can not say that the first sentences does not have semantics.  But whether
people buys into my vocabulary and build tools to support my vocabulary is the
problem of sharing but it is irrelevant to the questions of "semantics".  OWL
is just, as Jim said, another Ontology, aimed at promote the sharing of DL
terms.

In real life, we will use a lot of semantics that can not, at least not easily
expressed by OWL.  For instance, if I say

_:me foaf:firstName "Xiaoshu" .

Is there any OWL involved? No.  But is there any clear semantics, of course.

>
> Even so, why does it have to have new semantics?  The first thing first is to
> ground "things" on to the web.  Connecting a URI with another is as 
> easy as it
> gets, the same can not be said about connecting a particular row of one DB to
> another.  The added value is not "all" about semantics, it is 
> interoperability.
>
>
>
> [VK] Exactly my point! The added value is all about cross-linking, 
> which comes
> from the web infrastructure (URIs) and not from semantics!

To think "The added value is all about cross-linking" appears to be too 
extreme
to me.  SW technology acutually affect how to use a URI, don't you think?  It
is not that simple to attribute URI as the web infrastructure.  URI is
necessary but not sufficient, to which SW contributes, to promote
interoperability. The fact that SW is actually changing what and how we should
put under a URI, for instance, GRDDL and http-range14 issues etc. says
something about this.

I think John Madden's quoting of Jennifer Vendetti, Nothing. vs. 
Many-thing. is
a good analogy. System is larger than the sub of its parts.  If we view SW
mechanically, it adds "nothing" (or at least not much) to the CS field 
and web,
but as a whole, it adds lots.

Xiaoshu
Received on Monday, 3 April 2006 14:02:49 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 7 January 2015 14:52:25 UTC