Matlab and Mathematica would do much more careful validation, at least
I hope so. They're specialized math software, and they cost much more
than RB costs.
It's not as if RB was incapable of any calculation anyways. We're
talkin about 1 isolated case that should never happen if the
programmer has 2 ounces of common sense and programming skill. The
computer estimates it as best it can (in numeric form), and RB simply
passes on the CPU response, as expected, they have no reason to second
guess or validate the CPU in any way.
I still think this is a storm in a glass of water and that it is a
non-issue. I have seen much worse in my 5 years in computer
engineering school, and when it happens you work around it.
On Wed, 30 Mar 2005 20:12:53 -0600, Mark Lubratt <mark dot lubratt at indeq dot
com> wrote:
>
> On Mar 30, 2005, at 7:34 PM, Mathieu Langlois wrote:
>
> > I'm not trying to make excuses for the mathematical shortcomings, I'm
> > not saying the math results are correct, but I'm sure that RS has
> > better things to do with RB2005 than track down and compensate for
> > CPU/FPU shortcomings. If perfect math accuracy is THAT critical to
> > your projects, maybe you should consider another platform like Matlab
> > or Mathematica.
> >
>
> If it's CPU/FPU shortcomings, then Matlab and Mathematica would have
> similar problems. The issue is RS's decision in how they handle a
> result coming from the CPU/FPU.
>
> Try an integer division by zero in another programming environment (I
> tried gcc) -- you'll get a floating point exception or something
> similar.
>
> Mark
>
> _______________________________________________
> Unsubscribe or switch delivery mode:
> <http://www.realsoftware.com/support/listmanager/>
>
> Search the archives of this list here:
> <http://support.realsoftware.com/listarchives/lists.html>
>
_______________________________________________
Unsubscribe or switch delivery mode:
<http://www.realsoftware.com/support/listmanager/>
Search the archives of this list here:
<http://support.realsoftware.com/listarchives/lists.html>
|