Re: Negative zeros?

[Vic Norton]

I am receptive to all kinds of things, Przemek, but I have never been 
talking about anything other than the INTEGER 0. I was not talking 
about the real number 0, which is an equivalence class of Cauchy 
sequences or a Dedekind cut or what have you. I was not talking about 
the complex number 0, which is a pair of real 0's. And I was not 
talking about the limit of anything.

The 0's I had problems with were never computed. They were assigned. 
But because of the nature of my routine they were continually being 
multiplied by -1 (also an integer). If a 0 was multiplied an even 
number of times it stayed 0, and if a zero was multiplied an odd 
number of times it became -0. All of this was an artifact of my 
routine and had no meaning whatsoever. And the 0's and -0's popped up 
quite randomly in the output.

Thank heavens I was told about the x += 0 trick. Now we perhaps we 
can put this topic to rest.

Regards,

Vic

At 3:59 PM -0400 9/14/05, Przemek Klosowski wrote:
> OK, so as a mathematician you should be well receptive to Kahan's
> arguments, stemming from the complex analysis where there is a
> definite need to distinguish lim(x>0, x->0) from lim(x<0, x->0)
> because of branch cuts---otherwise, what would be the value of
> sqrt(i*x-1) in each case?


--
*---* mailto:address@hidden
|     Life is a jest; and all things show it.
|     I thought so once; and now I know it.
|           My Own Epitaph - John Gay, 1689-1732
*---* http://vic.norton.name



-------------------------------------------------------------
Octave is freely available under the terms of the GNU GPL.

Octave's home on the web:  http://www.octave.org
How to fund new projects:  http://www.octave.org/funding.html
Subscription information:  http://www.octave.org/archive.html
-------------------------------------------------------------