W3C home > Mailing lists > Public > www-webont-wg@w3.org > February 2002

Re: Next steps (Action: all)

From: Pat Hayes <phayes@ai.uwf.edu>
Date: Mon, 25 Feb 2002 19:31:05 -0600
Message-Id: <p0510147cb8a090cb6cc5@[65.212.118.219]>
To: Guus Schreiber <schreiber@swi.psy.uva.nl>
Cc: webont <www-webont-wg@w3.org>
>
>PLEASE NOTE:  there is an action item for all members at the end of this
>message (see *****)
>
>
>NEXT STEPS:
>
>We need to begin several processes in parallel if we are going to
>complete our work in time to get this language through the W3C
>process before the window of opportunity closes -
>
>   1. Language Features (functionality): We need to produce a document
>similar to the DAML+OIL Reference that describes the specifics of our
>language - starting place on this is determining the language
>features we need based on the requirements document and a
>"not-covered/not-used" analysis of D+O.

Let me strongly suggest that writing the document be done towards the 
end of the overall process, rather than in parallel. If this is done 
independently and in parallel, then we will waste far too much time 
keeping it in line with the other threads and arguing over priorities.

Also it would be good to have this driven by (4), ie people say what 
they want to be able to DO in OWL, and then we can argue about how 
exactly to provide features to let them do it. Arguing over features 
in the abstract tends to become aesthetics.

>   2. Implementation and Test Suite - we need examples that both show
>off the language and that can be used to test implementations (See 4.
>below)

Excellent idea.

>
>   3. Semantics: As evidenced by the layering discussion, developing
>the semantic model for the language (mandated by our charter) is not
>easy, but needs to be done.  We expect to produce both a model theory
>and an axiomization similar to the ones prepared for DAML+OIL.

I volunteer for this as my primary focus, but would again suggest 
that the MT be regarded as the central  activity, and the 
axiomatization provided subsequent to the MT.

OWL may well turn out to have a sufficiently subtle semantics that it 
will be difficult, if not impossible, to transcribe it in into a 
simple KIF-like logic without considerable extra work. (RDFS with 
datatyping is already in that state.) If that appears to be likely, 
it would be better to abandon the axiomatization rather than let this 
requirement warp the model theory. We have to be prepared to invent 
new things here, and if they don't fit exactly into traditional FOL, 
then tough.

>   4. Developing a set of methodological guidelines on how to use OWL
>in practice. This should show use of the language in handling common
>modelling issues. It can take the form of a walkthru, but it may be
>difficult to find one example domain that shows off everything. . The
>examples need to be realistic (and probably linked to the test cases,
>see 2). The guidelines should cover modelling issues for which no
>direct language feature is available, e.g. defaults, part-of
>relations.

Agreed this would be a very desireable product. It would be worth 
doing this even before the syntax is fully developed, using 
sketch-coding to give the general intent of the examples. I would 
like to see this kind of activity being rather like making a 
storyboard for a movie: a series of sketches that are not mandatory, 
but provide a central guide for the detailed technical effort.

>     These four must all "co-evolve" that is, we must work on them in
>parallel, but coordinate and make sure we stay consistent with each
>group checking the work of the others.

Oh sure, dead easy. Lets just all read everyone's emails :-)

Seriously though, it might be worth semi-formalizing a process here. 
If one subgroup is on the point of taking a decision that might 
impact what other groups are doing, they should try to emit a brief 
heads-up message asking for quick feedback of the form 
(can't-live-with-that/here's-why) from other groups. Then the issue, 
if one arises, can be focussed on by both groups intensively to try 
to find a way past the apparent deadlock, if indeed it is one. This 
might help to prevent a situation where one group invests so much 
effort that they are reluctant to reconsider a decision later.

Pat

-- 
---------------------------------------------------------------------
IHMC					(850)434 8903   home
40 South Alcaniz St.			(850)202 4416   office
Pensacola,  FL 32501			(850)202 4440   fax
phayes@ai.uwf.edu 
http://www.coginst.uwf.edu/~phayes
Received on Monday, 25 February 2002 20:31:10 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Monday, 7 December 2009 10:57:47 GMT