1/0=inf, -1/0=minf



On 8/5/2013 7:54 AM, Oliver Kullmann wrote:
> Also
>
> http://dl.acm.org/citation.cfm?id=1219095
> The rational numbers as an abstract data type
>
> could be of interest (it shows how to work well with 1/0 = 0).
I haven't studied this paper in detail, but since it describes an axiom 
system for rationals
in which 1/0 = 0  is a consequence, it appears that this is 
inappropriate for computer algebra
systems...  where applications seem to require that 1/0 is certainly not 
zero.

Perhaps this is another example of the gap between theory and practice 
in computing.

Richard.


>
> There are various papers follow up on this new algebraic structure.
>
> Oliver
>
>
> On Mon, Aug 05, 2013 at 07:23:40AM -0700, Richard Fateman wrote:
>> You might look at ..
>> computation with the extended rational numbers...
>>
>> www.cs.berkeley.edu/~fateman/papers/extrat.pdf?
>> (1994)
>>
>> RJF
>>
>>
>> On 8/5/2013 6:19 AM, Henry Baker wrote:
>>> I've been playing with rational numbers which allow the denominator to be zero.
.... snip