[DESIGN] Interactive text system.

2 messages in this thread from mud-dev2 in 2008-09

  1.   Robert Flesch <robertflesch@ya...com> 09-15 15:18
  2.   Mike Rozak <Mike@mx...au> 09-15 20:45

Robert Flesch <robertflesch@ya...com>

2008-09-15 15:18:49
Replies: Mike Rozak
We are looking to add an interactive text system to our virtual world.
We are going to use it to teach people how to bargain for items. Along
with other kinds of teaching.

For example.
The user would approach the bot. The bot can sense the person
approaching and say.
BOT - CONTEXT 1 - "Welcome, can I help you find something?"
USER - 1 -> n choices
BOT - looks up users response and if it finds an match based on the
opening context, sends the response back to the user. Then changes
context.
i.e. USER - "How much for this lamp"
     BOT  - "What is it worth to you?" - switch to context 2.0
OR
i.e. USER - "Lovely day today"
     BOT  - "Yes it is, and I have a hat that would look lovely on you"
     - switch to context 2.1

So you would end up with a deep contextual tree, that would either use
AI response from something like pandorabots, or go back to a "I dont
understand your question, please state it in a different way". (We
already use Pandora Bots as our bartenders, I have seem some hilarious
arguments between them and people who didnt realize they were bots.)

I had some ideas on how to write my own, but I suspect that this is a
well known subject in the MUD world. Can anyone point me to a paper, a
library(API), or some source code that could help me not reinvent the
wheel?

_______________________________________________
MUD-Dev2 mailing list
MUD-Dev2@mu...com
http://my.binhost.com/lists/listinfo/mud-dev2

Mike Rozak <Mike@mx...au>

2008-09-15 20:45:28
Robert Flesch wrote:
> We are looking to add an interactive text system to our virtual world.
> We are going to use it to teach people how to bargain for items. Along
> with other kinds of teaching.

>From your description, it's not clear what portion of the system you're 
looking for.

You might look at AIML. I'm using something like AIML in my game, except 
I've extended it to include probabilities, and be driven by code (not 
simplistic scripts).

You also need to figure out how furstrated players will get then the 
chatterbot doesn't understand what's being asked of it, and whether the 
benefits outweigh the frustrations. I found that open-ended chatterbot 
functionality, for the most part, was a net loss. I still use the flexible 
parser, but I always provide players with a list of the most common 
responses. 

_______________________________________________
MUD-Dev2 mailing list
MUD-Dev2@mu...com
http://my.binhost.com/lists/listinfo/mud-dev2

2 messages in this thread from mud-dev2 in 2008-09