Re: AI

From: Christian Loth (
Date: 11/15/00


On Wed, Nov 15, 2000 at 10:04:52AM -0500, Adam Hallam wrote:
> You could probably implement the "Eliza" AI system into mobs pretty easily,
> the AI bot itself has already been ported to run in *nix perl systems.
> The downside being that you would have to have the root install all of the
> perl libraries necessary.  But Eliza would definitely give the bots some
> semblence of a life.

No need for the perl overhead here, because the eliza algorithms are
pretty well known and easily implementable. The only thing is just, that
according to my experiences with such conversation systems, it won't just
be used! People don't talk to Eliza-NPCs. Simply because Eliza-Conversations
turn out very dumb after a while.

However, I have implemented a basic conversation system for the Gidayu MUD,
which works with a keyword/text mechanism, just like in the old Ultimas.
It also knows about keyword dependencies (e.g. A needs to be asked before
B, otherwise the NPC won't answer to B). This is nice for builtin quests
and similar cases. What is not yet implemented, but will be, are
in-game effects upon keyword triggering, like a spell being cast,
a reward, whatever. I will probably implement this via the dlsym
mechanism that is heavily in use with Gidayu. if e.g. the text
of a keyword starts with @, like in "@wizard_reward", the conversation
system tries to call the function wizard_reward from the .so file which
comes with the NPC.

The basic system is probably also very easily implemented for stock circle,
the in-game effects probably not.

Is a patch for the conversation system wanted?

- Chris

Christian Loth
Coder of 'Project Gidayu'
Computer Science Student, University of Dortmund -

     | Ensure that you have read the CircleMUD Mailing List FAQ:  |
     |  |

This archive was generated by hypermail 2b30 : 04/11/01 PDT