[Home] [Groups] - Message: [Prev in Group] [Next in Group]

nu.kanga.list.mud-dev

11798: RE: Distribution schemes (was Re: [MUD-Dev] Languages for MUD drivers)

[Full Header] [Plain Text]
From: "Ian Macintosh" <iman@issystems.co.nz>
Newsgroups: nu.kanga.list.mud-dev
Date: Sat, 20 Nov 1999 07:48:25 +1300
References: [1]
Organization: Kanga.Nu
    > Koster, Raph wrote:
    > Sent: Friday, 19 November 1999 05:00
    >
    > > From: Ian Macintosh [mailto:iman@issystems.co.nz]
    > > Sent: Wednesday, November 17, 1999 8:49 PM
    > >
    > > You have a few properties on mobs.  Thirst, Hunger,
    > Fear.  Now each
    > > mob will only obey one 'drive' at a time.  Each drive
    > gets a chance to
    > > calculate it's importance, and the one with the
    > highest value wins.
    > > That way, the thirstier an antelope gets, the closer
    > and closer it
    > > gets to the waterhole, even though it can see the
    > lions.  Finally, the
    > > lions nail one of the antelope, and the rest get to drink.
    > >
    > Stop now. :) We went down this road with Ultima Online.
    > With a model very
    > similar to what you describe (though we used food,
    > shelter, desire, taken
    > from Maslow).

My focus is very much the bottom rung of Maslow.  I don't plan go very
high either.  Anything higher than base motivations are *real* AI, and
I'm not aiming there.

    > Ask yourself whether anyone cares if an antelope falls
    > in the woods if there
    > is nobody there to see it. Then assess your design and
    > ask yourself some
    > tough questions about your objectives with the system.

The example I used was just generic to convey the idea.  You are quite
right re the importance, and I'm steering clear of that like the
plague.  I am just focusing on a more goal oriented, non-linear, event
driven approach to mob-progs than the traditional canned brain-dead
and puppet-like automation found everywhere else.  I am trying to aim
for an opponent with smarts.  About the highest I think I'll go on the
intelligence scale is a revenge motivation.  I do want mobs to learn a
bit from experience as well, via a simple reward/punishment filter.
If I find it un-workable in practice, I'll just lobotomise the
ancestor object :^)

I must admit that some of the drive I have in playing in this area is
just plain fun, nice to fiddle with, I enjoy it stuff.  On the cold
hearted business front, my only justification is seeing how hard I can
push the envelope.

Thanks for the advice :-)

Regards,

Ian.




_______________________________________________
MUD-Dev maillist  -  MUD-Dev@kanga.nu
http://www.kanga.nu/lists/listinfo/mud-dev