Re: A possible structure of the datatype system for OWL 2 (related to ISSUE-126)

On Jul 9, 2008, at 3:22 PM, Evan Wallace wrote:

> Let me go on record as opposing proposals that simple gloss over  
> the peculiarities of
> float and double and treat them as reals (thus I am agreeing with  
> Bijan's and Michael Smith's
> positions).  I would vastly prefer that we define our own real type  
> and use some explicit
> type coercion to compare values across these and other datatypes.
> -Evan

Hi Evan,

I think this is a reasonable position to take. I don't think it would  
effect my suggestion for defining the value spaces in OWL, nor the  
abandoning the assumption that a certain lexical type is uniquely  
associated with a specific value space.

Under Boris proposal, the modification would be to have a type like  
Integer, which is a subset of real, in this case, new types  
owl:Float  and owl:Double.

However, given that it has been expressed that this requires some  
work by developers, would it be possible to give us a case from your  
domain in which you would like OWL to behave differently when using a  
float value space as compared to a real value space?


Received on Wednesday, 9 July 2008 14:44:37 UTC