Sesame, OWL, and so forth

Once more I'm trying to collate my thoughts as to which Semantic Web backend I'm looking for. Not easy, as I've previously mentioned, and they do all seem to be in Java (bah).

However, I noted a few interesting Issues on the Sesame Issue Tracker: OWL support in Sesame; Enhanced inferencing using SWI prolog backend*; Make system configuration RDF-based; Aggregate functions in SeRQL (count, minimum, etc.). Seems other people want some of the things I do!

I've also noticed something about query languages. It seems there are two classes: end-point ones, which typically have a SQL-style syntax, and return values or other unstructured results, and ones that could support chaining more easily, where the query and its results are Semantic Web languages. E.g. RDQL:
SELECT * WHERE (?x, ?y, ?z); OWL-QL: a standard OWL knowledge base, with variables, wrapped up in a SOAP message. Perhaps due to different paradigms: one is intended to be a programmatic interface to the knowledge base (indeed, Jena's RDQL strings are an alternative to the programmatic form), while the other is meant to be a communication protocol for agents.

Of course, OWL-QL is a damn sight more powerful (though I like some of the features of SeRQL…).

Enough rambling!

* Interesting… Sesame currently just deduces the entire set of consequences from a KB, using that to answer queries. This proposal suggest using SWI Prolog as the backend to do backward-chaining.

Posted at 2004-04-22 07:51:44 by Richard • Link to Sesame, OWL, and s…
Comments, trackbacks.

SWI-Prolog

I'm messing around with SWI Prolog at the moment. I spent all morning and some of last night getting everything to compile, and getting it linked with C code (my ultimate aim is embedding it in a Cocoa daemon). Then I come to the most trivial thing — trying to use one of the in-built libraries.

Nowhere could I find how to do it. Eventually I stumble across the solution:

1. You have to be in the library directory at the time (in my case, /usr/local/lib/swipl-5.3.9/library).
2. Call use_module(library('semweb/rdf_db')).

Don't forget the closing full-stop! I always do, but then I haven't done much Prolog!.

Of course, now I have to do something useful with it. I might not end up using it at all — while it does do inference, I'd have to write my own Web-interfacing code (it can only load from files at the moment), and it doesn't have complete OWL support.

*sigh*

Why is there no perfect library?

Posted at 2004-04-22 06:35:38 by Richard • Link to SWI-Prolog
Comments, trackbacks.

New Forest Photos: Day Two

We kicked off day two with a visit to Exbury Gardens, a beautiful, well, garden full of rhododendrons, azaleas, and plants I don't know the names for. Oh, and daffodils.



There's a little steam train that runs around the gardens. Lisa's impish sense of humour soon had me risking life and limb for the sake of a good photo.



Later that day we went down to Lymington, on the south coast. A nice little town, where we had a spot of lunch, and a massive ice cream. I felt sick soon after this picture was taken.



Wonderful! I might post some pictures of animals later.

Posted at 2004-04-22 01:58:02 by Richard • Link to New Forest Photos:…
Comments, trackbacks.