Fourth Draft "Unsafe fp optimizations" project description.

[dewar@gnat.com]

<<Certainly not.  As further example, this entire message thread is
about how the proponents are going to make GCC ignore and violate IEEE
754 (and C99).  Not about whether they are going to do it, but in what
manner they are going to do it.  We already have a set of fast-math
transformations and they are going to add still more violations.  Do
you think those people are wrong?  Do you think you can stop them?
>>

I find the above simply bizarre, and certainly wish to disassociate
myself from this kind of extreme viewpoint!