W3C home > Mailing lists > Public > www-rdf-interest@w3.org > February 2004

RE: [semanticweb] Real World Semantic Web Tools?

From: Danny Ayers <danny666@virgilio.it>
Date: Fri, 13 Feb 2004 13:41:30 +0100
To: "Mansur Darlington" <ensmjd@bath.ac.uk>
Cc: <semanticweb@yahoogroups.com>, <kaw@swi.psy.uva.nl>, <www-rdf-interest@w3.org>
Message-ID: <BKELLDAGKABIOCHDFDBPMEIKFIAA.danny666@virgilio.it>


> These would include:
> 1) Ontology creation and lifecycle management (e.g. Protégé, OilED,
> OntoEdit, etc.)
> 2) An annotator for constructing mark-up documents from the corpus, or
> marked-up proxy documents (e.g. OntoMat, MnM, etc.).
> 3) A means for capturing and representing axioms or rules which
> formalize useful inferences in the domain (e.g. using the above).
> 4) An interface that invites the construction of queries (e.g. plugins
> for the above).
> 5) An accessible query engine which will handle querying and inferencing
> (using the axioms/rules) and present the result in a useful way.

...

> Fundamentally our questions are: Is it that the tools which we require
> are simply too immature (or dont yet exist) or that substantially more
> technical expertise is needed to use what is available than can be
> reasonably expected from information users?

It would be helpful if you could be more specific about the problems you're
encountering. Assuming I personally wanted to apply Semantic Web
technologies to a specific domain, at this point in time I'd probably cover
the points above in the following way:

1) Ontology creation - Protege with the OWL plugin
2) Annotator - it depends on the specific requirements, but I'd probably
have another look at Amaya/Annotea
3,4, 5 again it depends - it might be appropriate to use Protege with RACER
and custom forms/widgets, though more than likely I'd probably cover these
using Jena with a custom UI, using Jena's own inferencing capabilities and
RDQL support.

One part that is maybe lacking from the list is a form of persistence, but
there are a variety of triple store implementations available off-the-shelf.

Note that this is an open source (cheapskate) approach - an alternative
would be to buy a system like Network Inference's.

In the construction of the system there would undoubtedly be a need for more
specialised technical expertise than that of most information workers. But
this is exactly the same situation you are likely to face if you were basing
your solution around a relational database. The up front cost in expertise
may well be a lot greater at this point in time because of the relative
youth of the SemWeb technologies (RDBMS have maybe an extra 20 years). But
investment-wise, in the medium to long term I'd suggest there'd be a
considerable payoff.

Cheers,
Danny.

----

http://dannyayers.com
Received on Friday, 13 February 2004 07:50:26 UTC

This archive was generated by hypermail 2.3.1 : Wednesday, 5 February 2014 07:14:58 UTC