double rounding in x86

[Samuel A. Figueroa]

I have been corresponding with Mr. Tydeman about this.  I claim this behavior
is perfectly acceptable as far as the IEEE Standard is concerned.  He was
unable to produce an overwhelming argument to the contrary - innocent till
proven guilty.  Whether this behavior is desirable is another question.
However, due to the ambiguity in the IEEE Standard on this issue, I'd like
to submit a request for interpretation to the appropriate standards body.
I would appreciate it if someone could tell me who that would be, and what the
address is.  (The IEEE Standard has also been endorsed by ISO.)

- Sam Figueroa (figueroaacs.nyu.edu)