Re: ISSUE-5: n-ary datatypes - unimplemented incomplete design

Hi Jeremy,

I believe it is not true that there are no reasoners supporting n-ary
datatypes.  For example, RACER does. And it works rather well.

greetings,
 		Carsten


On Mon, 12 Nov 2007, Jeremy Carroll wrote:
>
> 1) unimplemented
>
> 2) no design for declaring n-ary datatypes
>
> To start with the second point:
>
> In the syntax doc
> http://www.w3.org/2007/OWL/wiki/Syntax#Data_Ranges
> data ranges with arity > 1 are introduced
> The only constructs used to produce such datatypes are a URI, or 
> DataComplementOf. It refers to the Semantics doc for defns of URIs, but only 
> 1-ary datatypes are listed there in.
>
> Essentially, the idea that there is a design for n-ary datatypes is a bluff.
>
> A design would include:
> - means for declaring such datatypes
> - some predefined datatypes
> - means for deciding on sensible selections of datatypes that work well 
> together.
>
>
>
> As far as I am aware, there are implementations that support many of the 
> features of the member submission - however, n-ary datatypes are not 
> supported by any od them. This increases my belief that n-ary datatypes still 
> requires active research work to take what is, I believe, a theoretical 
> design, to turn it into working code.
>
> I do not believe that this WG is the appropriate place for such research.
>
>
>
> Jeremy
>
>
>
>
>
>
>
>
>

--
*      Carsten Lutz, Institut f"ur Theoretische Informatik, TU Dresden       *
*     Office phone:++49 351 46339171   mailto:lutz@tcs.inf.tu-dresden.de     *

Received on Monday, 12 November 2007 21:51:29 UTC