[
Lists Home |
Date Index |
Thread Index
]
** Reply to message from "Michael Kay" <michael.h.kay@ntlworld.com> on Fri, 8
Nov 2002 10:07:37 -0000
> It also highlights the thing I haven't understood about RDF: in what way
> is it supposed to be different from the "binary relational" data models
> that have been lying around largely unimplemented for 30 years, except
> for academic prototypes and a few interesting niche applications in
> areas such as criminology?
Well, it's similar to a data warehousing application I was told about this
week, called Kalido
http://www.kalido.com/
which apparently uses a simple thing-association database model. It is
nonetheless quite performant, able to handle Shell's retail operations, and
Shell has more outlets than McDonalds!
The RDF model is a little different, being
subject-predicate-object
so the relationships (predicates) are one-directional, subject to object, but
it is similar enough. The topic map model is
thing-role-association-role-thing
where the associations are bidirectional (or multi-directional, as you can have
any number of roles in a topic map association), and I've seen relational
database developers independently come up with this solution for representing
complex and changing metadata links between data. So I would have to argue that
these models are not unimplemented, although I have the impression that at the
moment they are more implemented anonymously, as part of in-house relational
database solutions, than obviously, as part of off-the-shelf solutions.
Cheers,
Tony.
====
Anthony B. Coates, Information & Software Architect
mailto:abcoates@TheOffice.net
MDDL Editor (Market Data Definition Language)
http://www.mddl.org/
|