Re: Responses to "Draft of charter for NextWebOnt (Proposed) Working Group"

At 11:24 AM -0500 1/12/07, Kendall Clark wrote:
On Jan 12, 2007, at 10:52 AM, Jim Hendler wrote:

but let me be clear - I think there are two approaches that would be 
valid for the Working Group - either take usability and the real 
world into account, or leave the task of defining other OWL subsets 
to people who do.  What is a mistake is for the group to take it on 
and do it on purely theoretical grounds - all we'll end up with 
another travesty like OWL Lite


Jim,

1. Can you fix yr mail app so that it quotes other people's text 
properly? I'm interested in following this conversation, but it's 
kinda hard when all the voices blur together.

[JH] Hmm, looks fine in my mailer - I don't know what is causing the 
blur - I'll preface my remarks with "JH" for now, and if anyoe has a 
suggestion on how to fix, let me know



2. For my money (uh, literally!), DL-Lite and EL++ (and RDFS, in a 
different way) are already strongly and explicitly motivated by real 
world considerations. (And, FWIW, the existing TF doc *does*, inline, 
offer some 'real world' motivations for, say, EL++ and DL-Lite. These 
can be strengthened and should be.)


[jh] I've said often and in public that a small extension to RDFS is 
a major need.  Problem is if we want this to have widest impact, I 
believe we need to take into account more than just theoretical 
aspecxts.  I find that neither DL-Lite nor EL++ is currently 
answering the mail I get - they each have more in them than the early 
adopters, who are just learning RDFS and want mainly to do E/R level 
modeling, need.   All I am saying is I think we have to take 
experience in use into account and not have dozens of different 
subsets all with confusing names, and then claim victory.  I would 
like to see one "OWL Ultralite" that is as close to RDFS as possible 
- this is not my whim, it's what I've heard in presenting OWL to many 
potential customers and in working as a consultant for many people 
building OWL engines.   Tractability and rule-espressibility (i.e. an 
axiomization, not just a model theory) are important and so is 
"simplicity" of the OWL syntax generated (which needs to be close to 
RDFS syntax, IMO)



My new company is interested in the TF stuff specifically so that we 
can build a product around it, and take that product into the federal 
market. Perhaps playing the Expressivity Game against Oracle and IBM 
is a losing battle, but we don't know that (yet), and having a 
standards doc will help make that market, if it's going to be made at 
all.


[jH] No disagreement, we both agree this needs to be done, the 
question is should it be done in this WG, in the SWD group, in some 
new group, etc.   But I do believe that whatever group takes this on 
has to be responsive to user needs, not just theoretical concerns. 
Your perspective as someone who wants to sell this makes great sense 
to me - I am concerned because I got responses from people who were 
at OWLED that the tractible subsets document was what I needed, and I 
don't agree.  So if this WG feels that is enough, then I don't want 
it to be the one to have this responsibility - this is just my 
opinion, I don't state it as holy writ, but the request was for 
comments from potential players on the WG, and my participation will 
largely depend on what the deliverables are w/respect to the "low 
end" (aka dark side) which is where I think there is much more money 
to be made in the next few years...
  -JH
p.s. I believe the reason OWL Lite didn't end up being a usable 
fragment was exactly because the group argues as Uli did in her 
previous post, and I wasn't smart enough as chair to realize the 
fallacy of that argument (Frank van Harmelen told me I would regret 
that some day, he was right).

-- 
Prof James Hendler				hendler@cs.rpi.edu
Tetherless World Constellation Chair		http://www.cs.umd.edu/~hendler
Computer Science Dept			301-405-2696 (work)
Rensselaer Polytechnic Inst			301-405-6707 (Fax)
Troy, NY 12180