OWL: Experiences and Directions (OWLED)

[in progress; updated 25-03-2007]

In early to mid 2005, Bernardo had his thesis fairly well in hand and was looking to beef up his CV (and his general capabilities). I suggested that he should organize a workshop, specifically one for IWSC 2005. Service is part of being a good academic citizen, and having a nice little workshop on the CV really doesn't hurt. (Note, the main criterion of the workshop was that I wouldn't have to do anything. Yeah. Like that ever happens!) We decided to see if Ian and Peter would sign on as co-organizers as they have (had?!) a lot more OWL mojo/standing than us, and it gave a nice balance to the application (UMCP, Manchester, and Lucent). (Ian and Peter agreed with the proviso that they wouldn't have to do anything. Ha!)

The workshop was rejected by ISWC to our surprise and dismay. We held it anyway, right after ISWC. Thus, a workshop series was born!

The Philosophy

OWLED is a working workshop, not a presentation oriented one. How much we live up to this ideal is variable, but the point is to advance the state of the deployed art. We've had about 60 participants each year and conducted most of the business in the open, with everyone there (instead of breaking up into smaller groups). We have short (5 minute) and long (10 minute) presentations (with 5 minutes each for questions) with around 30 minutes per session for additional discussion. We also have had longer talks (a keynotey thing the first year for example), and in 2006 we had several extended sesssions to discuss OWL 1.1 and other standardization oriented issues.

I like OWLED to be a conversation between users, implementors, theorists, and other stakeholders. The main goal is to advance the state of OWL. Communication is important, but so is action. As much as possible, I like OWLED to work from consensus to produce substantial changes. Of course, not everything (or most things) can happen at OWLED (it is only two jam-packed days!), and not everything (or most things) can be done by groups. OWLED does serve as a big driver though. (For example, there was a substantial clean up of the OWL 1.1 documents in anticipation of OWLED2007, and Pellet support for OWL 1.1 was strongly driven by that impending event.)

Obviously, OWLED cannot channel the entire OWL community. But it has been a surprisingly coherent voice and I think a voice with a great deal of legitimacy. Obviously, it bears the stamp, both direct and indirect, of those who have organized it, and of the steering committee. But more it bears the stamp of those willing to do work. It's fine to come show off your stuff and try to attract users. That's a great thing to do. But if you want to get other people to do work (e.g., produce/support a W3C submission or implement your extensions in their reasoner) you have to do a lot of work. You have to show that its not so hard to implement and that it will, in fact, be widely used. Being an implementor of a serious system (or working with one) helps! Having funding (and labor) to share also helps.

(And for Pellet, I welcome source code!)

If something can help grow the community, people will tend to support it if there is a clearly sensible path to doing so.

While we've used the past two OWLEDs to drive (or try to drive) new work at the W3C, not everything needs to be formal standards oriented. For example, I do believe that alternative (and esp. natural language like alternative) syntaxes could be "effectively" standardized just by OWLED. Having a clear worked out spec and multiple implementations is the important bit.

The events themselves

OWLED the First

The first OWLED was held in 2005, in Galway Ireland and was co-located with IWSC 2005. Attendee count was a little indistinct since we had quite a few joint registrations with RuleML (some by accident; the registration site was rather messed up), but we had, I'd say, no less than 60 people in the room most of the time. Sometimes there were a bit more.

We ended up in an empty bar space, rather than in a more traditional meeting hall, which was funky, but worked out just fine.

OWLED the Second

35 submissions.

OWLED the Third (to come)

46 submissions.

Future events

Outcomes

The most obvious and significant outcome of OWLED is the OWL 1.1 language. The basic shape of the language was scoped out at OWLED2005, and the decision to try to get it to be a W3C member submission (and, we hope, eventually a recommendation) was made at OWLED2006. 

Similar events

By my suggestion, there is an OWL-S: Experiences and Directions. It's a little different (they were accepted by ESWC; don't they know that rejection build character!), and it will be interested to see how it plays out. 

I love the subtitle, "Experiences and Directions". It really captures the spirit of the event.