W3C home > Mailing lists > Public > www-archive@w3.org > December 2001

General Requirements Subgroup

From: <herman.ter.horst@philips.com>
Date: Tue, 11 Dec 2001 17:44:29 +0100
To: heflin@EECS.Lehigh.EDU, dlm@ksl.stanford.edu, phayes@ai.uwf.edu, jos.deroo.jd@belgium.agfa.com, jjc@hplb.hpl.hp.com, ned.smith@intel.com, Jim Hendler <hendler@cs.umd.edu>
Cc: www-archive@w3.org
Message-ID: <OF5C10FA4F.6988ED6C-ONC1256B1F.004F53E6@diamond.philips.com>
I would like to expand on two requirements already on the list under
construction, and discuss a new candidate for the list.

I think the following items are important and could be subsumed
under this heading:
- low learning barrier
- clear language concepts and meaning
- independence of the language concepts from syntax, XML serialization
(compare the basic specification document for RDF, which first presents
resources, properties, statements in a kind of set-theoretic way, 
from and combinable with various XML serializations) (this point also 
in some very recent e-mails to the Webont group by Jeremy and Frank)
- the user of the language should not need to know about reasoning 
strategies behind
the language
(compare programming languages: it is not attractive if a programmer would 
to know about LR-parsing and attribute grammars).

---Ontology querying
This requirement could be made more specific in the following direction.
The ontology language could be required to offer a semantic level of 
description of information on the Web that allows queries to be formulated 

independently from XML serialization, on the basis of the meaning 
of the information.  (So this links to the point already mentioned above 
user-friendly: independence from XML serialization.)
There is an analogy to the standard three-level architecture of databases: 

the ontology language could be viewed as delivering a conceptual level 
which frees information modelers and people wanting to pose queries from 
storage representation (i.e., serialization) details.

---Basic reasoning functionality
This requirement involves what is expected from an implementation of the
language.  What I mean here is certain basic reasoning functions which may 
assumed to come free to use by developers of 'intelligent' software 
making use of ontologies on the Web.  On the level of RDF Schema an 
could be that the meta-rule  'if the resource X is an instance of class C 
and class 
C is a subclass of class D, then X can be concluded to be an instance of 
does not need to be implemented by each software developer making use of 
documents involving RDF Schema.
This requirement, the presence of basic reasoning functionality, seems to 
be absent 
in the 13 requirements mentioned so far.
It seems to be appropriate to discuss this point (which has been mentioned 
in the 
Webont discussions already several times) simultaneously with the language 

Dr. Herman J. ter Horst
Philips Research Laboratories
Prof. Holstlaan 4, 5656 AA Eindhoven, The Netherlands

E-mail:    herman.ter.horst@philips.com
Tel:    + 31 40 27 42026
Received on Tuesday, 11 December 2001 11:47:56 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 22:31:39 UTC