W3C home > Mailing lists > Public > public-owl-wg@w3.org > July 2008

Re: ISSUE-53 Proposal to resolve

From: Alan Ruttenberg <alanruttenberg@gmail.com>
Date: Wed, 9 Jul 2008 01:49:33 +0100
Message-Id: <CCD97628-1BBD-440C-8395-D532157867A0@gmail.com>
Cc: Evan Wallace <ewallace@cme.nist.gov>, Ian Horrocks <ian.horrocks@comlab.ox.ac.uk>, "OWL 1.1" <public-owl-wg@w3.org>
To: Bijan Parsia <bparsia@cs.man.ac.uk>

Hi Bijan,

Closing the issue as withdrawn would satisfy my concern that the  
issues were not considered resolved. Thanks for the suggestion.

On the specifics you mention I concur that the issue of there being  
no proposal is moot. On the issue of fragility I'm not sure I agree.  
Specifically, the fragility associated with transitive and  
cardinality constraints is limited to a single property. The  
interaction in the case of n-ary's involves more than one property.  
This is at least different and I don't understand what the full  
implications would be. On the issue of whether dl-safe rules with  
built-ins are a better or reasonable alternative, I don't know  
because I don't remember discussing this as an alternative.

-Alan

On Jul 9, 2008, at 12:13 AM, Bijan Parsia wrote:

> On Jul 8, 2008, at 5:30 PM, Alan Ruttenberg wrote:
>
>> I'm content to close this issue, as long as doing so does not  
>> imply that the technical issues were resolved.
>
> What technical issues?
>
> Does it matter? I don't refer to this issue *at all* when working  
> on n-ary. It has no useful content.
>
> But fine:
>
> """" N-ary datatypes seem both fragile and wrongly expressive (as  
> compared, e.g., to dl-safe rules with built-ins)""""
>
> They are not fragile (or no worse than transitive roles). They are  
> rightly expressive (see use cases).
>
> (Some people still doubt the latter, but does this issue help  
> organize that? I'm sure Boris will raise it if it is necessary.)
>
> """there are no built-in n-ary datatypes, nor a language for  
> defining them. E.g., see:"""
>
> And that was just "it's a bluff" trash talking. We have a proposal  
> for built-in types and a language for defining them.
>
> There are issues with all this, but those are more narrowly  
> technical and are being handled anyway.
>
>> I'm confident we can find appropriate wording. It may the case  
>> that closing it as something other than "resolved" is appropriate.
>
> I propose we close this as withdrawn with a reference to Jeremy's  
> email.
>
> Close issue-53 as well along Ian's line.
>
> (Personally, I don't like stale issue and stale action items. At  
> some point, they can get in the way.)
>
> Cheers,
> Bijan.
Received on Wednesday, 9 July 2008 00:50:13 UTC

This archive was generated by hypermail 2.3.1 : Tuesday, 6 January 2015 21:42:05 UTC