Motion P1788/M0004.01_Keep_to_754 NO

[Alan Eliasen <eliasen@xxxxxxxxxxxxxx>]

   I vote NO for many of the same reasons mentioned by Guillaume
Melquiond and Arnold Neumaier.

   While I strongly believe that a standard that follows IEEE-754 as
closely as possible will make it much more likely that it is actually
implemented, it is not clear that confining ourselves to the existing
754-2008 standard will achieve a better interval standard than if we
analyzed wider possibilities or even suggested extensions to the IEEE
754-2008 standard or simplified implementations that are sufficient for
interval arithmetic.

   In addition, I believe that the wording of the motion and the
rationale statement are unclear and only confuse the issue, and don't
make it clear what this motion actually mandates or achieves.  I'm not
proposing an alternate wording, because I don't agree that a rewording
of the current motion is necessary nor productive for our discussion.

   When discussing specific terminology used in the 754 standard, (e.g.
"NaN",) we will understand each other better if we use the definitions
of the 754 standard, but this should not limit our analysis.

   Alan Eliasen