From: antispam@fricas.org   
      
   Michael S  wrote:   
   > On Sun, 14 Sep 2025 07:18:38 -0000 (UTC)   
   > Thomas Koenig  wrote:   
   >   
   >> Michael S  schrieb:   
   >> > On Sat, 13 Sep 2025 14:55:58 -0000 (UTC)   
   >   
   >> > 2. Does above said mean that caller has no way of modifying rounding   
   >> > mode used by callee? If true, it defeats one of original reasons for   
   >> > which Kahan invented rounding modes in the first place.   
   >>   
   >> The caller cannot change the callee's rounding mode without the   
   >> callee having been designed for this (by taking the rounding mode   
   >> as an extra argument and using IEEE_SET_ROUNDING_MODE itself).   
   >>   
   >> I think that's a good idea.  If a library routine is written   
   >> and debugged for a particular rounding mode, results should   
   >> not change because somebody up the call tree changed it.   
   >   
   > Personally, I never found the whole rounding modes business useful in   
   > my practice. So, can not say with straight face whether Fortran's take   
   > on it is good idea or not. But I can say with good level of certainty   
   > that William Kahan meant something else.   
      
   Kahan wrote a paper claiming that seting rounding mode to different   
   value is great debugging help, especially when no source is available.   
   So probably he really meant this.  But first thing that authors of   
   numerical packages learn is to set rounding mode to desired value.   
   There are some codes which should work reasonably for all rounding   
   modes, but a lot of code critically depends on rounding and will   
   not work with different rounding mode.  So, clearly this idea about   
   standard was wrong (as several other things in the standard).   
      
   --   
                                 Waldek Hebisch   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)