dynamic vs. static rounding

[Samuel A. Figueroa]

Now that I am working on an IEEE binding for Ada 9X, I reread some messages
I saved some time ago.  Among these messages, I found the following, written
by Peter Tang (I have deleted quite a bit of it):

  >From: uunet!antares.mcs.anl.gov!tangauunet.UU.NET
  >Date: Tue, 19 May 92 14:29:01 CDT
  >Message-Id: <9205191929.AA01243adingding.mcs.anl.gov>
  >To: numeric-interestavalidgh.com
  >Subject: dynamic vs. static rounding
  >
  >All supports of IEEE rounding controls (on current machines) that
  >I know of is dynamic. That is, the machine instructions round
  >according to a control word that can be changed in runtime.
  >Although some may insist that dynamic control is mandated by the
  >Standard, some have argued otherwise. Along this line, one can
  >also ask whether static rounding control suffice in
  >applications.
  >Questions:
  >1. Does the Standard really mandante dynamic rounding?

I reread the Standard, and I thought dynamic rounding modes were not required.
But Robert Dewar suggested I look at the definition of the word "mode" in
section 2.  It says that a mode is "a variable that a user may set, sense,
save, and restore to control the execution of subsequent arithmetic opera-
tions....  The following mode shall be implemented: rounding, to control the
direction of rounding errors."  So I can't imagine how one could have only
static rounding control, and still be able to claim to fully implement the
Standard.