[MUD-Dev] Dynamically changing room descriptions

[Ola Fosheim Grøstad <olag@ifi.uio.no>]

Travis Casey wrote:
> My idea, then, is to give each element in the description an
> importance rating.  The description could then be manipulated so that
> the most important things are described first.

[...]

> This seems like a fairly obvious idea, now that I've thought about
> it... has anyone else suggested it, and/or is anyone already using it
> in their mud?

Dunno if anyone have suggested it.  But I did decide that I was going to
assign importance ratings to objects when I started paying attention to
text-to-speech synthesis.  Basically, I think it is required if you want to
avoid overload.  As I have just started thinking about MUDs from this angle,
it is all kinda fuzzy to me right now, but I'd basically vote for the basic
thinking you are arguing for.

I'm considering fuzzy logic or fuzzy numbers.  Maybe I then could do some
simple, but clever, computations on this as well?

Anyway, the issues involved in generating descriptions sound somewhat
familiar to problems research on story generation will have had to deal
with.  Maybe they are?  The trouble with MUD design is that there are so
incredibly many fields that COULD be relevant.  For instance there was an
announcement about an "intelligent agents" conference in sci.virtual-worlds
recently.  That could be relevant too.  Unfortunately I get to see all these
interesting announcements, but they are usually not very good at making the
papers available on web after the conference.  Many papers are probably out
there though, we just have to find them!

I think it might be a good idea if list members were willing to volunteer to
search the web for relevant information and then made them available to the
list.  It is too bad that (most of us) discuss these topics with very little
knowledge about what has been done of research in the past decades.  Not to
imply that such discussions are bad, but they could become even better. 
Basically, if I were to program monster AI, I would rather look to AI and
psych for ideas than to existing MUDs.  However, discussing AI research in a
MUD context would be very interesting.

So really, the next time any of you find something that catch your
interest.  Why not go all the way and do a real search for info, then toss
it to the list?

--
Ola
(Admittedly most of my own attempts at providing such information have not
been very good. I'll try to do better :-)



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