[Cfp-interest 1810] Re: Exact subnormal results

[David Hough CFP]

> If underflow (and overflow) need this level of explanation, and I felt
> like information overload on that, do we have a serous problem in the
> context of C2X and the above subject?

In 1977,
I don't think anybody was worrying about how 754 arithmetic would be grafted
onto language definitions that hoped to also encompass non-754 arithmetic.

So what does C range error mean for non-754?    A small result, certainly,
but does it matter if exact or inexact?     How does before or after rounding
underflow detection in 754 systems affect the C definition of range error?

> Also, I noticed talk of C23.  Has C2X morphed into a fixed year slot now?

On the conference today, it was mentioned that the C committee has decided
to refer to C2X as C23 from now on.     They're being hopeful.