W3C home > Mailing lists > Public > www-webont-wg@w3.org > January 2003

Re: HP reactions to AS&S and OWL

From: Jeremy Carroll <jjc@hplb.hpl.hp.com>
Date: Wed, 15 Jan 2003 09:32:12 +0000
Message-ID: <3E252A9C.2050500@hpl.hp.com>
To: Jim Hendler <hendler@cs.umd.edu>
CC: Jeremy Carroll <jjc@hplb.hpl.hp.com>, webont <www-webont-wg@w3.org>

Personally I have a lot of sympathy with the views expressed here; I am 
unclear whether my colleagues will find this line of argument sufficiently 
compelling that we have got OWL sufficiently right.

I think as representative I have a procedural problem, how best to ensure 
that HP has enough time to come to a thought out position, rather than a 
substantive problem; I think HP *may* have a substantive problem.

The counterpoints to Jim's points here are:
1: Euler does not set out to be OWL Lite complete - if it turns out that it 
is, that would be significant - this suggests a CR exit criteria which 
includes a completeness proof for Euler (or similar tool) with respect to 
OWL Lite.

2: the ease of exposition argument does not show that we got the semantics 
right. The view of my colleagues is that OWL Lite as the basis for 
interoperability means that there should be semantic interoperability - 
this means that many tool sets etc. should be implementing OWL Lite 
semantics, and lots of different incomplete OWL Lites will not deliver 
interoperability.

Timeline wise I don't believe HP can resolve this before the planned last 
call period (certainly not before 23rd for the last call vote).

I am beginning to fear that I may even end up having to vote against my own 
document!!

Jeremy



Jim Hendler wrote:

>>> I'm confused Jeremy - we saw ample examples of implementation of OWL 
>>> tools at the f2f, and I'm still not sure what features of Lite you 
>>> believe are unimplemented.  I asked several times at the f2f for 
>>> people to bring up things they think are as yet unimplemented but 
>>> needed for moving to PR, and very few were mentioned - my lab took 
>>> two actions to produce some of this.  You brought up none and 
>>> volunteered none.  I have been drafting some starts at the 
>>> implementation experience, and I don't see any major holes -- please 
>>> identify any you have so we can start to fill them
>>
>>
>>
>> Currently noone has a complete OWL Lite reasoner.
>> I have every reason to believe that NI will deliver one, but that's 
>> one rather than two.
>>
>> If OWL Lite is meant to be easy, and a trustworthy basis for 
>> interoperability then we should be looking at more than one complete 
>> OWL Lite reasoner before exiting CR.
> 
> 
> do you mean only Lite?  Does Euler fail any of the Lite tests at this 
> point?  I was under the impression from something Jos said that he 
> handled virtually all the Lite tests - did I misunderstand?
> 
>>
>> Also I am dubious that full tools that incompletely support lite will 
>> give the sort of interoperability that wg members were looking for.
>>
>> So to justify its existence Lite not only needs to be implemented, but 
>> we should get the feeling that almost all OWL tools will migrate to 
>> having complete OWL Lite capability. My colleagues are more doubtful 
>> of this than I am.
> 
> 
> First - I understand that you have a tools focus, and will ask the 
> following about tools:
>  I don't understand the "...almost all OWL tools will might to having 
> COMPLETE OWL Lite capability" -- I would expect almost all OWL tools to 
> support the bulk of the Owl Lite functionality, but would be very 
> surprised if very many of them are complete.  This is because I expect 
> OWL Lite to be the sublanguage of choice to be supported by 
> NON-REASONING tools (like many of those we saw demoed at the f2f) - for 
> example, almost all the tools converting from DAML do only some subset 
> of Lite - but that is because they are for markup, crawling, parsing, 
> querying, etc.  etc -- and none of those promise to have a COMPLETE 
> reasoner, since most only use simple reasoning to make the tools more 
> powerful.
> 
> I believe the above motivates LITE just as much, if not more, than the 
> sort of migration you discuss.
> 
> However,  probably more importantly - I actually never thought OWL Lite 
> was as important for tools as it is for Documents - most of the existing 
> DAML ontologies are in or near the OWL Lite coverage -- and thus I 
> expect most ontologies for the near future to be in Lite. Teaching 
> students to build Lite ontologies is easier, for me, than teaching them 
> the more complex DL constructions, and gives them access to a lot of 
> power.  Ian often stated that the only motivation for Lite was ease of 
> implementation, but many of us never agreed with that motivation - and 
> felt ease of exposition and learning to be important as well.
> 
> So let's not be totally tool-centric, from a document-centric 
> perspective (and what is the web if not document-centric?) there's a lot 
> of evidence that people will use Owl Lite.
> 
> (to confirm the above, [1] is a list of every OWL language construct and 
> which ontologies use them.  Playing with that you can find that most of 
> them are only using stuff from Lite.  The documents themselves may be 
> missing some assertions needed to make them stickered Owl Lite 
> documents, but those can be added in conversion - can't fault the DAML 
> documents for not generating to an abstract syntax that didn't exist 
> when they were written.)
> [1] http://www.daml.org/ontologies/features
> 
> 
>>
>> I volunteered no tools because within the time frame requested it is 
>> unlikely that HP will be able to deliver any. That does not mean that 
>> we are not working on OWL. If the wind blows in our favour we just 
>> might have a version of Jena with OWL support in it, (i.e. a 
>> programming environment which processes imports and allows RDF/XML to 
>> be manipulated using the abstractions from OWL).
> 
> 
> that would be great, but I do understand that you may not be able to do 
> this on time - I didn't mean to criticize you for not volunteering, I 
> just meant that at the time you didn't pose that there were outstanding 
> implementational items needed.
> 
>>
>> If you can circulate a draft of the implementation report I would like
>> to ask my colleagues for their comments.
> 
> 
> it certainly won't be ready for circulation until after we go to LC. At 
> this point I'm gathering data, but all my WG time is spent working 
> documents, agendae, etc. until we get the LC documents out the door.
> 
>>
>>
>> Jeremy
> 
> 
> 
Received on Wednesday, 15 January 2003 05:07:51 GMT

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