Spike Jones wrote:
> Mike Lorrey wrote:
> > [battling automatons] would necessitate the development of a logical
> > system
> > of discerning vulnerable points and components, as well as strategies
> > for using hazards against the opponent while minimizing exposure of
> > one's self to them. This would be a true programmers competition...
>
> Now that is a hell of an idea, Mike. I wonder if there is any way to
> carry out such a robot battle entirely in a simulation, with enough
> fidelity that any robot sim could be built in hardware and the performance
> would be similar to what the software version predicted. Doing it
> in software would way reduce the cost of developing these things.
> Then the champions on the computer would get built and let them
> duke it out for cheering hordes of proletariat. spike
I remember a DOS program, back in the '80s, by the name of "PCROBOTS"
that did more or less that. It took simple C source files, with a few
custom routines to drive sensors and cannons and motors, and loaded
them into virtual robots and ran them against each other. It wasn't
the only one of its kind by any means, but it's the one I remember most
since coding a good 'bot for that was one of my first non-trivial C
programs. I would not be surprised to find it still exists today,
though there are probably more complex (more realistic) simulations out
there.
One thing keeping this particular example from BattleBots or similar
competitions: the only good weapon was a cannon, and BattleBots
specifically prohibits untethered projectiles.
This archive was generated by hypermail 2b30 : Fri Oct 12 2001 - 14:39:59 MDT