I'm not saying every polynomial multiplication program can be shown to
be correct; just the method I suggested happens to be pretty simple.
If you write a polynomial multiplication program that has certain
breakpoints, e.g. switching to a different method like Karatsuba or
FFT at size 3000, then
On Tue, May 26, 2009 at 3:55 PM, rjf wrote:
>
> Disregarding the obvious " marketing fluff" of Wolfram's statement, it
> seems to me plausible that some programs are more likely to be correct
> (at least in well-understood basic components) than some theorems'
> proofs. Here's why.
>
> A program
Disregarding the obvious " marketing fluff" of Wolfram's statement, it
seems to me plausible that some programs are more likely to be correct
(at least in well-understood basic components) than some theorems'
proofs. Here's why.
A program can be run on test data.
A program can sometimes be analy
On Tue, May 26, 2009 at 10:48 AM, Dr. David Kirkby
wrote:
>
> There was an discussion on sci.math.symbolic, which like many on
> newsgroups, often go off the subject into something more intersting (or
> more boring)!
>
> It was started by me under the title "Wolfram Alpha claims to be a
> primary
