Thinking Clearly .

Following 6 weeks after the release of 1.5-RC1, we’ve released Pellet 1.5.0. Notable changes since the 1.4 release series include incremental reasoning, axiom tracing, and an optimized classifier.

Thanks to excellent user feedback, we were able to resolve many issues identified in the release candidate. The most significant enhancements since that release are updates to newer versions of OWLAPI and Jena, and enhancements to the Jena interface to support incremental reasoning. Full descriptions of the changes since RC1 are accessible using the milestone view of the Pellet ticket list.

For more details on the release, see the full release announcement in the pellet-users archive.

For future releases, we’re enhancing the rules support and are being supported to implement incremental classification. If you are interested in influencing Pellet’s future, or are just an active user, join us on the pellet-users mailing list and tell us what you think.

We’ve been collaborating with the National Cancer Institute this year to improve Pellet’s performance when working with the NCI Thesaurus, a very large OWL ontology for which the interesting problems are TBox problems.

In a previous post, we highlighted our success reducing classification times for the Thesaurus by a few orders of magnitude. Together with that work, we also added explanation services to the NCI infrastructure. These improvements have had the desired effect—our client has reported that the improvements enabled some long-standing modeling bugs to be diagnosed and repaired.

More recently we’ve undertaken an effort to add incremental classification support to Pellet, with the goal of enabling Thesaurus editors to see the impact of their changes in real-time. The first step towards incremental classification is partitioning of the ontology into modules. For us, this meant a review of the existing state-of-the-art in ontology modularity and some engineering to turn prototype algorithms into professional software. To date, it has yielded some very satisfying results. Though we never waited long enough for it to complete, our estimates are that the first-cut partitioning code was taking about 50 hours to partition the latest version of the Thesaurus—after a few days of software engineering, the partitioning takes about 5 minutes.

The impact this has on classification times is something we expect to detail in a future post.

For now, we’re taking it as another example of what Bijan said here recently: the performance story in OWL is getting better all the time. As you can tell, we’re happy and proud to be a part of that.

When we announced our 2007 internship program, Semantic Summer, we didn’t expect to receive more applications than we had open slots. But that’s what happened, and we’re pretty happy about it.

We accepted 5 applicants:

1. Ron Alford is working this summer on improving the DL Safe Rules support in Pellet; and on OWLSight—our unreleased Web 2.0-style ontology tool: it’s GWT in front of Pellet, essentially. Ron starts his MS at UMD in the fall.
2. Christian Halaschek-Wiener is working on incremental reasoning in Pellet this summer with Mike Smith and Evren. Christian is finishing his PhD at UMD this fall.
3. Pavel Klinov is working on probabilistic reasoning in Pellet with Bijan, Evren, and Mike. Pavel is a PhD student at University of Cincinnati.
4. Vlad Kolovski is working on policy management in Pellet. Vlad’s also finsihing his PhD at UMD.
5. Petr Křemen will be working in the fall on SPARQL-DL query evaluation in Pellet. Petr’s presently a grad student and researcher at Czech Technical University in Prague.

It’s a UMD-heavy list, since that’s where most of us come from, but we’re especially happy to have some European computer scientists joining us, too. We’re very bullish about the future of Pellet, OWL, Description Logics, and semantic technology generaly in the EU.

So many workshops and conferences….

Anyway, I’m one of the organizers for SWWS ‘07, so I thought I’d give it a shout out. There is a lot going on in this space, but it still seems to be the case that the common knowledge (and thus unified message) of what’s going on isn’t there. SWWS seems a good place to try to pull that together.

While scalability of the underlying engines (e.g., query engine, parser, reasoner) is important, equally important is scalability of the UI. It would be also really great to get a better handle of what kind of applications face what sorts of scale problems. I guess that Linked Data folks have something to say here.

Why? Cause it’s the bomb.

I’m definitely a come-lately. I mean, I was fairly old school “oh no! inaccessible, insecure, hideous Javscript web pages!” I don’t remember when I stopped turning Javascript off in browsers. I certainly found many of the Javascript based applications (e.g., various Web based mail apps, Google maps fer sure, some reasoner interfaces etc.) useful, compelling, and much more friendly than an Java applet or webstart (though not always!) and certainly better than Firefox/Mozilla XUL based apps (i.e., the experience didn’t seem substantively worse, it was sometimes better, and it had better installation/cross-browser behavior).

But my heart never warmed to all this. It still seemed better for Web applications rather than Web content—I certainly didn’t like having to use Javascript rather than some elements or CSS to do simple things like hide/show, more/less text. Exhibit touched me, but I ended up punting on S5 for a while, in spite of some early success, because of printing and math issues. I was intrigued by the use of Javascript libraries to bring HTML 5 features to legacy browsers. We could be living in the future!

And it was a bit of Javascript mediated future that has won me over…the inexpressibly wonderful jsMath library.

The only way I can really indicate how cool I find the prospect of being able to enter equations user LaTeX and have them render properly in any browser with clean resizing and the source intact is to say that finding jsMath completely displaced all my iPhone lust. All of it. I’m not saying I’d turn down an iPhone if you handed me one, but jsMath has the same sort of elegance.

I can have math in the browser! And it works! Without upgrading the browser or fighting MathML.

It’s not living in the future; it’s living in a better past.