[WF-Scripting] AI scripting

[Bryce Harrington]

On Tue, 6 Mar 2001, RaFaL Pocztarski wrote:
> Bryce Harrington wrote:
> 
> > Man, wouldn't that be great!!  Yes, that's exactly what we want to see
> > happen.  I think we've got half a chance, too.
> 
> I think that's only a metter of time. I don't really think that WF could have
> any serious commercial rival after, let's say, 10 years from now. I just don't
> know if it would be 10 or maybe 2 years.

Let's hope for 2 years.  ;-)

> > *Nod*  Having this as the goal will give good purpose to the work.
> > I think you'll find it not *too* difficult to get this set up.  Good
> > luck!  Lotsa people can help so if you get stuck, ask around.
> 
> OK, I will when I know what to ask about, but now I don't know what's really
> wrong. :)

How are things coming now, btw?

> > Who can say at this point.  Acorn has taken approximately a year,
> > and while Mason is more sophisticated, we also have a headstart due to
> > having already done Acorn.  In the end it really depends on how many
> > people can contribute time to it.
> 
> I'm just wondering how much time I will have to learn and write some useful
> scripts. Of course it's never too late to contribute but I'd like to have some
> of my scripts in Mason's first playable version. 

Well, one of the first really big requirements that needs to be in place
before we can come anywhere close to making predictions about when Mason
will be ready is the STAGE server.  I would think it would require the
better part of a year, at least, before STAGE is usable anywhere close
to where Mason will need it.  This of course is highly dependant on the
amount of work people can put into the server over that time.  The good
news is that right now the team appears to be growing, and individuals
who had run out of time previously, seem to have found extra time to put
into it.  So, figure something on the order of a year.  Plenty of time,
no?  So don't worry if for some months you've got little time; put in
what you can when you can, and we'll get there.  :-)

> Btw, what scripts would Mason need first?

Zzorn is a better one to answer this than I, but I would assume that a
safe starting point would be wildlife scripts - deer, flocks of birds,
etc.  I suggest this in large part because people have already been
writing these sorts of scripts for Cyphesis and Acorn, and so there is a
wealth of examples there.

One thing that would be a big help, is to look at this page:
   http://www.worldforge.org/website/rules/mason/tasks/
and help Zzorn generate a section to list scripts-needed.  I think this
would be a good starting point for the scripting, and would help us
scope out what exactly we're going to need others to help with.

> > Yup, but I think you'll enjoy it.  :-)  You don't need to look at all
> > the code - just the parts that involve the pigs and other critters, and
> > if I remember correctly this is all in its own subdir, so you won't be
> > overwhelmed by a lot of code (unless you want to!)
> 
> Yes, the code is divided in a very nice way. When I figure out all I need it
> should be easy to experiment a litle bit.
> 
> > Cool.  I think you have a good approach.  If you're limited in time,
> > focus first on the interesting bits, and the tutorial can come
> > secondarily if you have time.
> 
> First of all I have to just run the software, I thought that it would be much
> simplier. Then I think starting to make some really simple scripts wouldn't
> take me very long. At that point I should be able to write some simple beginers
> tutorial. I have a lot of work now but this project is much more interresting,
> so when I start I would always find some time to play with the scripts.

How are you coming with this?  Let us know if you've run into problems
so that we can get you in contact with the appropriate people to get ya
fixed up.

> How the scripts will be executed in a real game? Thousands of animals
> and NPC's would need a lot of computing power. Will everything be done
> on the server?

Initially we'll do it all on the main server.  However, our architecture
allows us to establish separate machines for running AI independently of
the main server machine.  I expect we're going to explore a few
different mechanisms for handling this.  If you can come up with clever
low-processor means for computing the action of, say, a whole herd of
sheep, that'd be a really cool thing to have.

-----------------------------------------------------------------------
Bryce Harrington     bryce @ neptune.net    bryceharrington @ yahoo.com