At 01:56 AM 4/27/2001 +0200, Tom Schwaller wrote:
>good point. One could detect this kind of typo with test cases, but
>testing if the method really works as expected would mean
>to run the test a least for a day. Is this really a good idea?
I think it's a great idea. Provided it doesn't suck up serious CPU cycles,
I could even find a Linux box for it to run on. We could have simultaneous
tests running for various periods: 1 day, 1 week, 1 month, 3 months, 6m, 12m.
We might not get up to 3 months and beyond due to reboots...