W3C home > Mailing lists > Public > public-rif-wg@w3.org > February 2009

Re: notes from OWL and RIF datatype coordination meeting

From: Bijan Parsia <bparsia@cs.man.ac.uk>
Date: Mon, 16 Feb 2009 18:58:11 +0000
Message-Id: <518271BC-46BC-493E-97BA-908136B762E7@cs.man.ac.uk>
Cc: Alan Ruttenberg <alanruttenberg@gmail.com>, public-rif-wg@w3.org, public-owl-wg@w3.org
To: Ivan Herman <ivan@w3.org>
On 16 Feb 2009, at 17:48, Ivan Herman wrote:

> Just an additional issue, though
>
> Alan Ruttenberg wrote:
> [snip]
>> In order for OWL to be precisely defined we need to base our
>> specification on the mathematical properties of numbers, not the
>> properties of implementations.

Er...In order for OWL to be precisely defined we need to precise  
define OWL. Full stop.

> Boris has reported major implementation issues regarding the  
> current setting. This is independent of the RIF coordination issue.  
> AFAIK, (and I may be wrong, sorry if that is the case) C&P also had  
> implementation issues. I definitely had problem on the RL  
> implementation side if I based it on an existing RDF environment.
>
> Ie, this may be one of those cases when implementation concerns may  
> have to have a priority.

The mathematical structure that has floats disjoint from decimals is  
perfectly well behaved and fairly standard. There is a sensible  
mathematical structure in which rationals are disjoint from integers  
(e.g., let the integers be 0, plus the closure of 0 under the  
successor operator  s and the minus sign (where -0 = 0); let the  
rationals be pairs <x, y> where x is an integer and y is an integer ! 
= 0; under this definition, no rational is an integer)

> As I said on the RIF/OWL call, I also have serious concerns with  
> any RIF OWL incompatibilities, because there are applications that  
> use both rule systems (ie, RIF in future, hopefully) and OWL at the  
> same time. Those applications will have major issues.
>
> Ie: as far as I am concerned, it is not that clear cut...

Well, for me, it is clear cut...pro disjointness :) DISJOINTNESS  
FOREVER!!!!

The other point is also clear: The "mathematical properties" of a  
type called "float" esp. in relation to a type called "decimal" are  
not univocal. It depends on what we *want*.

Personally, I think, and have always thought, that there is both user  
and implementation value in making them disjoint. I was proven right  
about implementation... ;)

Cheers,
Bijan.
Received on Monday, 16 February 2009 18:54:42 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Tuesday, 2 June 2009 18:34:03 GMT