Re: Zeroes and infinities

[David Hough 754R work <754r@xxxxxxxxxxx>]

I think the discussions of zeros and infinities and pow have all served
to underline the observation that no matter what default is prescribed 
for the exceptional situations in which they arise, 
somebody will take exception to
it (because it's wrong for his application).     That's one reason they
are called exceptions.

So correct exception handling necessarily encompasses

1) default handling that best serves most applications most of the time

2) portable facilities for alternate exception handling for all the other 
   applications and times



I'm not surprised that 1) continues to generate a lot of discussion
since there are elements of judgment involved in some of the decisions.

754 punted on 2) by specifying optional traps that are useless in portable
programs.    I'm surprised at the resistance expressed since 2001 to attempts
to begin to address that in 754R.     I'd understand the resistance better
if it were expressed in terms of needs of providers and users of mathematical
software.

The only motivation I've had for working on 754R for seven years has been to
remedy the gaps in 754 about alternate exception handling and expression
evaluation (which really is alternate exception handling too).   Ratifying
existing practices for fma, correctly rounded base conversion, etc. should not 
have been worth more than about a year's effort. 

The other worthwhile innovation in 754R is fixing the semantics of decimal
floating-point operations.    I'd view that as filling a gap in 854.
But it's not an omission that has affect me much professionally, so I leave
it to others who are more directly concerned.