Re: Mutilple maps?

[Stan Shebs <shebs at apple dot com>]

Lincoln Peters wrote:
> 
> On Tue, 2002-06-18 at 09:24, Stan Shebs wrote:
> > >From hard experience, I should say that this approach, while
> > having the appeal of greater abstraction in the code, doesn't
> > work very well in practice.  What typically happens that the
> > abstraction is easy to understand and implement, you do it, but
> > then the real purpose of the abstraction is hard to implement and
> > never gets done.  Xconq is littered with hundreds of these - clouds
> > for instance, which have map layers, GDL, etc, but there isn't any
> > way for a game to use them (I originally had the idea of using them
> > to protect parts of a battlefield from air attack and maybe ranged
> > fire, but that part was never done).
> 
> It seems to me that the best way to implement clouds might be to use a
> terrain coating that behaves like a particle system, using position,
> velocity, acceleration, and gravity.  It could then affect affect units
> using thickness, attack-terrain-effect, and/or defend-terrain-effect.
> It could even be flexible enough that the same algorithm could be
> adapted to nebulae in space games, or forest fires in games that deal
> with guerrilla warfare (using attrition to damage units).
> 
> I might have enough time this summer to work on such a project.

Clouds should have both bottoms and tops of different elevations,
at least if you want to do the aircraft thing - you can have
fighters attack ground forces if the cloud bottoms are a couple
hundred meters up, but high-level bombers are blocked.

A practical difficulty with all this is how to display it to the
player; cloud density is doable with stipples, at the cost of
readability, but tops/bottoms are harder.  One way might be to
display the net effect of clouds - gray out cells that a
high-level bomber cannot attack (of course then you have a bit
of footwork to let the bomber pass over potential victims in
the same cell, only reveal them if the clouds blow away).

Stan