Re: status of xsd:duration in OWL (and RIF and SPARQL) - ACTION-164: RDF WG

On 4 May 2012, at 12:08, Ivan Herman wrote:

> 
> On May 4, 2012, at 12:10 , Bijan Parsia wrote:
> 
>> On 4 May 2012, at 11:05, Ivan Herman wrote:
>> 
>>> (Removing some mailing lists from the CC list, because that seems to be a purely OWL, and possibly RIF WG discussion...)
>>> 
>>> I am not commenting on the technical issue. The current status of the WG is based on the latest extension of its charter announced in:
>>> 
>>> https://lists.w3.org/Archives/Member/w3c-ac-members/2011OctDec/0066.html
>> 
>> Can we use a public reference.
> 
> I am sorry; charter extensions and administrations are always member confidential... I cannot change that. 

Ok, but all it says it allows is changing a reference to normative. So, "fixing things" is already an expansion on the strict reading, though obviously, it'd be untenable not to fix things.

> But the situation is also described (less formally) on the home page of the WG:
> 
> http://www.w3.org/2007/OWL/wiki/OWL_Working_Group
> 
> which refers to the dependencies on the XSD specification. Updating the dependencies does not mean adding new features, ie, new datatypes.

Pfft. This is just sheerly a judgement call. Are we adding a new feature (a NEW DATATYPE) or updating a dependency (the XSD set of datatypes).

If XSD had broken a datatype from our perspective, would we just remove it,or would we "add" it back into the "owl" set of datatypes?

If one we didn't include because it was broken by our perspective were fixed, what's wrong with adding it back in (given that "using all the XSD datatypes that are feasible" sorta IS the feature)?


>>> in my reading that does not allow adding new datatypes, or change the semantics of existing ones (except if there are technical issues created by the final version of the XSD document compared to the version we used when OWL 2 was finalized).
>> 
>> Technically, I don't see where that's explicitly authorized. I can't quote the member only email to demonstrate this. A narrow reading would forbid even that. A broad reading could allow various tweaks.
>> 
>> IOW, if we want to fudge it we can :)
> 
> There is a fairly clear text, I believe, in the 
> 
> http://www.w3.org/2005/10/Process-20051014/tr#rec-modify
> 
> which clearly states what an edited recommendation is. Based on the alternatives in 7.6.2., although conformance may be affected by possible changes, no new features are allowed for that.

what?
7.6.2 Classes of Changes to a Recommendation

4. New features
The first two classes of change require no technical review of the proposed changes, although a Working Group may issue a Call for Review. The modified Recommendation is published according to the Team's requirements, including Publication Rules [PUB31].

New features are explicitly allowed. So we don't even have to get into what "feature" means. It's explicitly allowed.

> Adding a new datatype would mean issuing an OWL 2.1 (or OWL 2.01:-). 

Clearly not, afaict. We could do that, of course. We would have to have a nominal LC, CR, and PR, though these are frankly, to my mind, ridiculous.

>>> Doing so would not mean an 'edited' OWL 2 document, but a new version instead. That would require rechartering the OWL WG or creating a new group.
>> 
>> 
>> A quick rechartering wouldn't be so hard, I think, for this narrow point. I'm happy to start thinking about a new group (it's getting toward time).
>> 
>> However, I personally hope that the W3C wouldn't be so hidebound and rule entangled to prevent a useful tweak on narrow procedural grounds, esp. when there's a reasonably expansive reading available.
> 
> I think we are bound to the rules here.

Well, which interpretation of the rules :)

> If we decided to bend them significantly, we may get public/member push back and discussion. Do we want to go there?
> 
> To make it clear: I am not against, per se, adding new datatypes to OWL. But, I believe, we have to go through due process of, essentially, creating a new group with everything that it requires and means...

On the flip side, perhaps we should just put a CG together and make a submission. As long as it's de facto adopted, who cares otherwise. Jean-Pierre, are you going to be an OWLED? I'll try to get some momentum going there.

Cheers,
Bijan.