[MUD-Dev2] [Design] Non-cliche content creation

[Damion Schubert]

On Jan 25, 2008 4:31 AM, cruise <cruise at casual-tempest.net> wrote:

> Thus spake Damion Schubert...
> > On Jan 4, 2008 9:23 AM, cruise <cruise at casual-tempest.net> wrote:
> >> My current r&d effort (outside of metaplace) is creating a system to
> >> produce npcs that can interact and spontaneously produce quests and
> >> missions - from the simple "I need 10 loaves" to "kill my rival for the
> >> throne".
> >>
> >
> > My contrarian thoughts:
> >
> > 1) The number one problem with interesting quest/NPC generation is the
> cost
> > in development and QA of content generation.
>
> "The number one problem with 3d graphic engines is the cost in
> development and QA of content generation."
>
> I'm not disagreeing, merely pointing out that it's somewhat irrelevant.
> Yes, it's difficult and time consuming. So are a lot fo things if you
> want them done well. Once they're done, however, they stay done. Once I
> have a quest generation system, all my quests from ever on are free.
> Sounds like a good trade off to me :P


This depends on whether or not you figure out how to make your randomly
created
content both engaging and interesting.  Is it possible?  Maybe.  It's worth
noting,
though, that the randomly created content in both Diablo (map) and
Daggerfall
(maps and quests) were pretty miserably bad.

Let's twist it and look at it from another angle.  Remove other people from
the
equation.  Remove interactivity altogether.  Do you think you can write an
algorithm that produces an excellent fantasy novel that matches Tolkien
(or hell, for that matter, that even matches Robert Jordan and Piers
Anthony?)
Of course not - great art requires more than following a formula, it
requires
emotional resonance, artful foreshadowing, solid chemistry between
interesting
characters... stuff that humans still are vastly superior than computers at.

Now, one may argue that the quests that you currently see in WoW are
by no means Tolkienesque.  I would argue that they are still vastly
superior to the madlib style randomly generated quests I most often see
when people talk about procedural story content for MMOs.


> These won't be "quests" in the formal manner that currently exists in
> most RPG style games - the very nature of the generation system will
> mean that they're closer to informal requests, and there's no set way to
> complete them. If you get asked by someone to kill the king, you're free
> to ignore them, turn them in to the guards, pass the request on to
> another character (NPC or player), etc. All of these will have
> ramifications, obviously.
>
> Turning the would-be regicidal NPC in might cause the guards to ask
> other players to investigate this individual, in turn uncovering a whole
> secret group dedicated to overthrowing the current government, who then
> start attacking publicly, bringing more players in on either side.
>
> My current sticking point is getting NPCs to be able to believably value
>  abstract things, like the reliability of an assassin :P
>

If I understand what you're designing, then it seems to me that another
large concern is ensuring that the world makes sense to the players.
Players can only see one part of the elephant - that which they have
direct contact with, and it makes it harder for them to get a real sense
of what's going on.

As an example, there was a lot of talk a while ago about creating 'dynamic
ecologies' - the dragon eats sheep, and if you kill all the sheep, the
dragon
attacks the village.  The problem is that this cause and effect isn't
necessarily
visible - all a newbie knows is that yesterday, the village was safe, but
today,
he got wtfpwned as soon as he logged in by something he clearly can't
handle.  Static, predictable worlds have a huge, hidden advantage that
most people ignore - the fact that they are static and predictable.  Players
can actually make sense of what's going on around them, and if they
log off for a week, things will still make sense when they log back on.

Having large-scale world change works best when there is relatively few
large decision points that are visibly accessible to the player.  A good
example is territorial control games, which are shown on an easily
accessible political map (Mythic does stuff like this).

--d