Two more AI bugs
Stan Shebs
shebs@apple.com
Mon Jul 8 12:33:00 GMT 2002
Lincoln Peters wrote:
>
> On Fri, 2002-07-05 at 11:32, Jim Kingdon wrote:
> > Now, there still is a need for end to end tests (e.g. must cross the
> > strait in less than 20 turns), but testing each unit reduces the
> > amount of work you need to try to do in the end to end tests.
>
> You know, it seems that every Xconq game I've ever written unearthed at
> least one bug in the AI. Perhaps what we need is more people writing
> game modules.
Or more people working on the AI code. :-) I estimate the current
generic AI code is about 1/10 of the size that it should be, to
be effective on a variety of games. That's one of the motivations
for multiple AI types that can be tied to specific games; it's
really hard to make a single body of code that can understand the
consequences of all the tricky things that a person can write into
a game design.
Stan
More information about the Xconq7
mailing list