Positive and negative zero: Paper from Prof. William Kahan

[behoffski <behoffski@...>]

G'day,

Regarding Steve Donovan's query (Anzac Day, 25 April 2014):

> What is -0 and should I be worried about it?

Prof. William Kahan is one of the most influential people in numerical
methods, ranging from calculators such as the HP models from the 70s,
through to being a main architect of Intel's floating point
infrastructure (hardware and software design, including compiler and
language considerations), which followed on to the IEEE 754
floating-point standard for binary floating-point arithmetic:

        http://www.cs.berkeley.edu/~wkahan/

One of his papers specifically addresses the need for a signed zero:

        Branch Cuts for Complex Elementary Functions, or Much Ado About Nothing's Sign Bit
            in The State of the Art in Numerical Analysis, (eds. Iserles and Powell), Clarendon Press, Oxford, 1987.

This paper is now available as a PDF on the web in various places, such
as:

        http://people.freebsd.org/~das/kahan86branch.pdf

Also, there is a Wikipedia page about signed zero:

        http://en.wikipedia.org/wiki/Signed_zero

cheers,

behoffski (Brenton Hoff)
Programmer, Grouse Software