Fourth Draft "Unsafe fp optimizations" project description.

[dewar@gnat.com]

<<Of course the whole issue here is the -ffast-math option, which indeed
gives up on such accuracy requirements to achieve better speed, so
your comment is not relevant here.  Think of this as the Ada relaxed math mode.
>>

Well the question is whether the built-in routines can be used in normal
mode. I would say that it makes sense to say yes on this. After all, IEEE
has no notion of required accuracy for elementary functions in the first
place.