Re: CR Exit Criteria

On 21 May 2009, at 00:53, Peter F.Patel-Schneider wrote:

> I've copied the current CR Exit Criteria here so that my changes  
> can be
> more easily discussed in email.
>
> I think that these criteria are very close to suitable.
>
> My comments are on lines starting with >
>
> peter
>
>
>
> ******************************
> The goal of the CR phase is to demonstrate the existence of multiple
> interoperable and practically useful OWL 2 systems. As a minimum,  
> the WG
> suggest that the following conditions should be met.
>
> 1. Resolve dependencies on rdf:text (currently at Last Call) and XSD
>    1.1 Datatypes (currently at Candidate Recommendation).
>> or put rdf:text as risk as described in previous email

We agreed to make it "at risk" in the spec, so we should mention it  
here. I added "Note that rdf:text is marked as "at risk", and may be  
removed from the OWL 2 specification."
>
> 2. Two different implementations of each of the following:
> * An OWL 2 DL entailment checker
> * An OWL 2 EL entailment checker that takes advantage of some
>   of the claimed benefits of the EL profile
> * An OWL 2 QL entailment checker that takes advantage of some
>   of the claimed benefits of the QL profile
> * An OWL 2 RL entailment checker that takes advantage of some
>   of the claimed benefits of the RL profile
>
>> The benefits appear to me to be the ones applicable to  
>> implementations
>> in the profile's preamble in Profiles, including
>> EL - good scaling, polynomial time
>> QL - LOGSPACE data complexity, SQL rewriting,
>> RL - RDFS++ implemenation, rule-based implementation
>> I'm not suggesting that the benefits be listed in the exit criteria,
>> but it might be a good idea to say that they are listed in  
>> Profiles at
>> the start of each profile's section.
>
>> For EL, QL, and RL I think that there should only be a need for two
>> different implementations, with at least one taking advantage of the
>> profile's benefits.

Agreed.


>
> 3. Two different implementations of an OWL 2 Full entailment checker
>    implementing useful subsets of OWL Full and taking advantage of at
>    least some of the claimed benefits of OWL 2 Full†

I favour going with something like "and passing at least some of the  
non-DL test cases". We could combine the two formulations and have  
"implementing useful subsets of OWL Full and passing a useful subset  
of the non-DL test cases".

Of course the only downside of this kind of loose wording is that  
there could be arguments later as to whether they have been satisfied  
or not.

Ian




>
>> I'm having a bit of a problem finding where these benefits are  
>> listed.
>> I suppose that they could include full upward semantic compatability
>> with RDFS, which would give rise to the option below.
>
> 4. For each of the standard OWL DL test cases, at least two
>    implementations that pass the test and which claim to be conformant
>    OWL 2 DL entailment checkers
>
> Note: Some of these criteria depend on the OWL 2 test suite, which is
> expected to continue to evolve. For the purposes of these criteria, we
> will only consider "Approved" tests which are not "Extra-Credit", and
> which were approved before some cut-off date, to be determined later,
> some time during CR.
>
> † Another option here would be to make this last phrase more test
> oriented, e.g., "and passing at least some of the non-DL test cases"
> ******************************
>

Received on Friday, 22 May 2009 11:05:29 UTC