Re: Scope

On Wed, 7 Nov 2001, Dan Brickley wrote:

> One way to get started on some detail might be for people to focus on
> what they've built, prototyped etc to date in this area using RDF.
> Setting aside for now the question of 'what is RDF _query_ versus
> _rules_', and swapping some brief summaries of what tools, techniques
> we've used for doing things like asking questions of an RDF
> database/service or representing and using inference rules over RDF
> content. We can worry about whether to call it a database or a
> knowledgebase, a query language or a logic system etc etc later.
> First, I'd love to know what people have been building...

I do not need more of an invitation than that to start advertising ;)

We have been working on a system called Sesame, which implements the
query language RQL. RQL, originally developed by the people at ICS-FORTH
in Greece, is a language with an OQL-like syntax that focuses on path
expressions through RDF graphs, and the intermixing of schema and data
information in a single query. RQL offers native support for RDF Schema
inferencing (such as class and property subsumption). An RQL query
typically uses a select-from-where syntax, where the select clause acts
as an output filter, the from clause is used for path expressions
through the graph, and the where clause can be used to express boolean
constraints on the variables bound in the from clause. The result set of
an RQL query consists of a set (a Bag, if you will) of resources and/or
literals. We have currently not implemented data typing. One of the
things that we are currently looking at is support for multiple
languages (specified with the xml:lang attribute).

The idea of the Sesame architecture is to make it portable across
different types of storages. The query engine is not running directly on
a database, but on a "Repository Abstraction Layer" which is a Java API
that connects query engine to database (in our test case, we are using
PostgresSQL, but we are working on a more generic SQL2 implementation).

There is a demo server of Sesame running online at
http://sesame.aidministrator.nl/, you can also find additional
documentation there.

An interesting side note might be that we plan to use this system (or
actually, have already started using it) in the case studies of the
On-To-Knowledge project. Sesame will act as a mediating repository for
different tools, such as a text mining tool that produces RDF, an
ontology editor, and several types of graphical user interfaces.  The
aim is to help make a corporate intranet's information more accessible.

Regards,

Jeen
-- 
jeen.broekstra@aidministrator.nl
aidministrator nederland bv - http://www.aidministrator.nl/
julianaplein 14b, 3817 cs amersfoort, the netherlands
tel. +31-(0)33-4659987, fax. +31-(0)33-4659987

Received on Friday, 9 November 2001 04:30:26 UTC