Re: Issue-114

On Jun 30, 2008, at 11:07 PM, Alan Ruttenberg wrote:

> On Jun 30, 2008, at 5:25 PM, Bijan Parsia wrote:
>> Sure there has.
>>
>> 	1) To handle more RDF graphs. (class/property punning is in OWL  
>> Full, with exactly the same semantics.)
>
> Could we see some examples of this usage in the wild? I have been  
> asked

Where, for what purpose? AFAIK, whether it appears in the wild has  
nothing to do with that issue.

> that re: the ID-based reification and am in the process of  
> collecting examples.

Similarly, it has nothing to do with this issue.

>> 	2) To handle various integration/alignment situations.
>
> Could you please be more specific about what these situations are?  
> If we are going to base decisions on such I would like to have  
> something that we can think about.

It's extremely frustrating to have this same conversation over and  
over again. We had *exactly* the same conversation over data/object  
property punning *more than once*. Each time, you reset the clock to  
zero, i.e., claimed that either no use cases were given or they we  
vague.

>> Furthermore, these are already in and in OWL2 implementations.
>
> So are there bugs. Shall we include those in the spec as well?

I don't know what this was intended to do, other than be a bit of  
unhelpful snark. Do I really have to explain to you how actual  
implementation weighs in? How the lack of user complaints weighs in?  
(Including the point that that might indicate that no one is using  
the feature?)

>>> It may be bad practice, but is is requested. To the extent that  
>>> we are chartered to provide features that have been identified by  
>>> users as widely needed,
>>
>> Actually, I don't see that we're chartered to do that:
>
>> """The mission of the OWL Working Group, part of the Semantic Web  
>> Activity, is to produce a W3C Recommendation that refines and  
>> extends OWL. The proposed extensions are a small set that:"""
>
> Interesting point to cut the quote. Note the similarity of point 1  
> in what immediately follows to what I wrote (no accident that ;-)

That was the *right* point to cut. we're chartered to refine and  
extend OWL. There are proposed extensions, but the language does not  
require that our refinements and extensions adhere to the same critera.

>> 1. have been identified by users as widely needed, and
>> 2. have been identified by tool implementers as reasonable and  
>> feasible extensions to current tools.
>
>> The starting point is OWL 1.1. I think we should deviate from  
>> included features that have been interoperably implemented only  
>> when there are clear and strong arguments against them.
>
> I have never understood this point of view.

You've never understood the view that if we have interoperable  
implementations of a previously specced feature then we should drop  
those features only if there are clear and strong arguments against  
them?

I'm flabberghasted. It seems like a simple and straightforward point  
of view. In fact, it seems like a basic principle of standardization,  
i.e., to try to unify existing behavior.

> The OWL 1.1 submission was one of several starting points.

Perhaps we don't speak the same idiolect. I find in the charter:
	"The starting point for the Working Group is the OWL 1.1 member  
submission."

> It is the working group that defines the features of OWL 2,

Am I not a member of the working group? Did I not write "I think"?  
Did I not add extra criteria?

> without any presumption to default acceptance of any of the inputs.


Are you speaking as chair or as advocate?

Cheers,
Bijan.

Received on Tuesday, 1 July 2008 06:57:05 UTC