W3C home > Mailing lists > Public > public-esw-thes@w3.org > January 2005

'constraints' on the SKOS Core model

From: Miles, AJ (Alistair) <A.J.Miles@rl.ac.uk>
Date: Mon, 24 Jan 2005 14:18:56 -0000
Message-ID: <F5839D944C66C049BDB45F4C1E3DF89D18DB50@exchange31.fed.cclrc.ac.uk>
To: "'public-esw-thes@w3.org'" <public-esw-thes@w3.org>

Tom wrote:
7. In sections such as Annotation Constraints, the text says
   "a concept may have no more than one definition per
   language".  This seems a touch too prescriptive, as if
   the Guide were declaring constraints for an XML schema
   instead of providing guidance in an Open World context.
   Maybe "should"?

Mark wrote:
The introduction explains how examples, rules and constraints are
depicted throughout the document. For a constraint, I'm not sure I
exactly understand what it means. Is it to give information on the
intended use of a vocabulary item?

The idea behind describing 'constraints' was to support stronger validation
of SKOS Core instance data, where that is required.  A 'constraint' being a
restriction on the graph structures allowed in instance data which if
violated would raise a validation error. 

One of the strongest business cases for XML and XML schema (or DTD etc.) is
the possibility for strong validation of document structures (which for XML
equates to data structures).  Although SKOS Core is a semantic web
application, which implies a commitment to open-world style usage, I believe
SKOS Core should also be a viable solution for closed world (i.e. enterprise
internal or business transaction) applications, where strong data validation
is necessary.  Many people in my experience  find this sort of use case
(where SKOS Core is essentially a specification of a data exchange format)
compelling and more relevant to their needs.  SKOS Core competes well with
pure XML formats here, because of the ease with which the specification can
be 'extended/refined' (more on this in another email).  

So anyway, I thought we should support those people who want to do stronger
validation of instance data by describing constraints on the SKOS Core data
model.

However, I think Tom's wording change suggestion is fine.  I.e. we describe
a set of constraints, which 'should' be implemented, but obviously this is
RDF so at the end of the day you can do what you want.

Does this sound sensible?

Cheers,

Al.

 



---
Alistair Miles
Research Associate
CCLRC - Rutherford Appleton Laboratory
Building R1 Room 1.60
Fermi Avenue
Chilton
Didcot
Oxfordshire OX11 0QX
United Kingdom
Email:        a.j.miles@rl.ac.uk
Tel: +44 (0)1235 445440
Received on Monday, 24 January 2005 14:19:30 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:38:53 GMT