software testing (was)

Brandon Van Every vanevery at 3DProgrammer.com
Tue Feb 4 03:02:30 CET 2003


Andy Freeman wrote:
> "Brandon Van Every" <vanevery at 3DProgrammer.com> wrote in message
> news:<UUf%9.4052$6P2.467701 at newsread1.prod.itd.earthlink.net>...
>> Andy Freeman wrote:
>>> "Brandon Van Every" <vanevery at 3DProgrammer.com> wrote in message
>>> news:<5N3%9.3161$6P2.354427 at newsread1.prod.itd.earthlink.net>...
>>> Yup, and that's an argument for flexiblity and a whole lot of
>>> testing.
>>> Testing time is strongly related to "how long does it take to get
>>> working code".
>>
>> Testing, yes.  Flexibility, why?
>
> Because when the point of testing is to reveal problems that you're
> going to fix.  That's where flexibility comes into play.

To me, the point of a test case is to prove that the incremental
functionality I've just implemented actually works.  I keep the test case
around and run it occasionally when I make other changes in the app, to
prove that I haven't broken anything by making a change elsewhere.  But
frankly, since I'm the only one designing this thing and I'm my own Code
Nazi, there's no risk of spaghetti intercommunication.  I simply don't build
things like that.

>>> Huh?  Why does your AI depend on the topology of your world?
>>
>> You ever tried to write one?
>
> I've written several.  Figuring out where things "want" to move
> isn't all that dependent on topology (or even the "gravity" or
> motive system).

Written a game AI on a sphere?

--
Cheers,                         www.3DProgrammer.com
Brandon Van Every               Seattle, WA

20% of the world is real.
80% is gobbledygook we make up inside our own heads.





More information about the Python-list mailing list