An alternative view for Owl 1.1?
I was totally frustrated that I was unable to stay at ISWC this year -unfortunately politics and science often go hand and hand, and I found myself in the vortex (no, not those politics, but I must admit it’s the first time in a long time I’ve enjoyed watching election returns come in). Anyway, I wish I could have stayed around for OWLEd, because reading the blogs, notes and papers I find myself very frustrated - not that I think anything is wrong with the work, and I particularly like some of the technical innovations that make the OWL Full/OWL DL distinction largely disappear from a users’ perspective (a great thing in my opinion). The problem is that I don’t think what will help OWL get more use in the real world is twiddling at the top end — what we really need, and I have been working on with a number of people, is more twiddling near the bottom — the gap from RDFS to OWL Light is way too big, and there’s lots of room in the middle for a much, much simpler OWL Minimal (bad name, but connotes the idea - real names could be OWL Fast, OWL Prime, or RDFS++).
In the paper that David Wang gave at ISWC (I really wanted to be there!!!) he described what he found when he surveyed a lot of things out there - similar papers by Tim Finin and others have also explored this space - and one very heavily used space is that where RDFS plus some minimal OWL is used to enhance data mappings or to develop simple schemas. FOAF, for example,would not be nearly as used without the owl:inverseFunctionalProperty on the email fields - which allows FOAF crawlers to have a property that unifies the different Foaf files created by different sites in a useful way. SKOS went to great lengths to avoid any use of OWL, as it gets redesigned this OWL Minimal looks to be perfect for its needs. The tools for this level are also coming, descriptions by IBM and Oracle seem to imply some RDFS/OWL support moving to the Database level - and htis makes sense - this minimal level of OWL can remain polynomial - plays the Database game by the database rules (and got published in VLDB)
In short, either as part of OWL 1.1 or opposed to it (depending on the poilitics) we should have some people working on a clean subset of OWL that is easy to learn, easy to support, and provides useful new functionality at the data schema/controlled vocabulary level — I think this would increase OWL use and uptake, and would provide a natural path from RDF to OWL through this middle layer - providing more room (and thus more money to be made) in the OWL space.
Ora Lassila made a proposal similar to this to the W3C AC meeting in Edinburgh, and was encouraged to pursue it, he and I are working on producing some sort of description and semantics (axiomatic - because this should fit nicely on a business rules base) — I should admit it is probably my foot-dragging that is causing the problem now - if I ever manage to write this thing up, it could get done. Dean Allemang and I are doing an OWL book which will have this level described as well - and it turns out to be pretty natural. In fact, if one takes some user considerations into effect (by not including some features that are easy to compute but hard to do in OWL syntax) it looks like it is a very clean level w/respect to OWL syntax - one can avoid needing to make users understand the restriction statement. btw, I should mention that the original version of OWL Lite proposed by Frank van Harmelen was much closer to what I describe than the travesty that the WG eventually produced - Frank is a smart guy, we should have listened to him!!
I’ll be happy to share more of our thoughts on this and would love some help if there’s someone out there motivated to help on this who is also good at HTML editing for the W3C formats…
-Jim Hendler
p.s. And I should mention now, early, and in public, that if the “charter” being circulated for OWL 1.1 were to come to a W3C I would oppose it on the grounds that moving away from the RDF syntax is a non-starter with me - but that’s a topic for some later blog…
