Re: Is this a numerical problem?

[Jordi Gutiérrez Hermoso]

On 19 August 2010 10:17, Ozzy Lash <address@hidden> wrote:
> 2010/8/19 Jordi Gutiérrez Hermoso <address@hidden>:
>>
>> there is quite possibly at least a cosmetic bug here in
>> how Octave is representing a number, namely, it doesn't display the
>> same thing Matlab does.
>
> I think, reading the original e-mail, that he is getting the same
> result with Matlab R2009a.

I believe the inconsistency is how 0.999995 is displayed in each with
"format short". Octave displays 0.99999 and Matlab displays 1.0000,
from what I can read in the bug report.

This may well be a gcc issue. I've encountered other people who have
been unhappy with how gcc's stdlib handles rounding for some numbers
such as these.