Thread Links Date Links
Thread Prev Thread Next Thread Index Date Prev Date Next Date Index

Re: A proposal for the next motion



> Date: Wed, 13 May 2009 13:41:03 +0200
> From: Vincent Lefevre <vincent@xxxxxxxxxx>
> To: stds-1788@xxxxxxxxxxxxxxxxx
> Subject: Re: A proposal for the next motion
> 
> On 2009-05-13 10:03:05 +0100, John Pryce wrote:
> > (2) An implementation shall be called 754-conforming if: (i) its  
> > underlying system is 754-conforming (Definition D3) and if the  
> > implementation supports intervals of each of the five basic 754 floating 
>                                        ^^^^
> > point formats: binary32, binary64, binary128, decimal64, and decimal128; 
> > or (ii) it is functionally indistinguishable from case (i).
> 
> Do you really mean "each" here? In particular, I don't think that
> supporting both binary *and* decimal is always useful in practice.
> And I think that the IEEE 754 requirements concerning the formats
> are sufficient, and 1788 shouldn't demand more than IEEE 754.
> 
> -- 
> Vincent Lefèvre <vincent@xxxxxxxxxx> - Web: <http://www.vinc17.org/>

	Vincent,

	I think what must have been intended here is the interval
	support of each of the basic floating-point formats also
	supported by the underlying 754 environment.

	There is language further down discussing variable precision
	formats as optional in much the same way.  I also presume
	them to be supported if they are supported as floating-point
	types in the underlying environment.

	Or so I believe...


				Dan