Re: Jump lines as roads.

[Eric McDonald <mcdonald@phy.cmich.edu>]

On Mon, 23 Aug 2004 mskala@ansuz.sooke.bc.ca wrote:

> I did exactly that in a game I was working on.  One problem I found was
> that the pathfinding code didn't seem to be smart enough to take full
> advantage of the fast routes - resulting in a big advantage for human
> players if they were willing to spend the time moving the units a step or
> two at a time.  It probably would have worked better if the fast routes
> were, or were close to, straight lines.

This is one of the few advantages we lost in May of this year when 
we reverted back from Peter's Astar pathfinder implementation to 
the simpler one that Xconq presently uses (and used before around 
October or November of 2003). With Astar, one has an evaluator 
(cost function) that is applied along various paths, and the best 
path is chosen through use of the evaluator. (Provided you have a 
good evaluator, of course.) The evaluator is free to consider 
whatever it wants, and ours did take into account such things as 
movement price when weighing the cost of a particular path.

Eric