Re: Motion P1788.1/M004.01
On Tue, 10 May 2016 13:17:40 -0400, Lee Winter wrote:
> No. There are no unbounded real intervals. Every single computable
> interval over rational FP numbers is bounded.
False. Entire is truly unbounded, and is a required entity in 1788.1.
It is true that 1788.1 does not mandate two-output division, but it
certainly does not disallow it, and this operation returns two intervals
that may be truly unbounded on one side, when the divisor includes zero,
which is a true zero.
Remember that although the bounds of intervals are restricted to
representable FP numbers, the interval actually describes ALL real
numbers that fall between these bounds, plus the bounds themselves
if they are finite.
Michel.
---Sent: 2016-05-13 14:26:45 UTC