On Fri, 12 Nov 2010 13:09:07 -0500, Simon Slavin <[email protected]>
wrote:
> On 12 Nov 2010, at 4:25pm, Dr. David Kirkby wrote:
>
>> I think in general mathematical software like Sage is very difficult to
>> test.
>
> Yes, I read your response and agree with this. But it's worth pointing
> out that commercial maths applications like MatLab /do/ have testing
> like that. That's one of the reasons they're so expensive. I'm sure
> they don't test every possible value, but they are thorough and they are
> run for every point version.
I think I know what you mean, and I have no wish to be snide hereby; yet
resist the obvious, I cannot:
If considering the set of real numbers, wouldn’t testing “every possible
value” require at least א -sub-1 tests? With א -sub-1 tests betwixt each
test? Ad infinitum, literally?
Someone please answer this urgent question, as I am bad at math and I’d
like to get to the bottom of it.
free(OT); return 0; }
Samuel Adam <[email protected]>
763 Montgomery Road
Hillsborough, NJ 08844-1304
United States
http://certifound.com/
([0] …with extraneous spaces because I doubt everybody’s MUA supports BIDI
overrides properly.)
_______________________________________________
sqlite-users mailing list
[email protected]
http://sqlite.org:8080/cgi-bin/mailman/listinfo/sqlite-users
