johnbue at msn.com
Thu Aug 2 15:54:32 New Zealand Standard Time 2001
Daniel Harman writes:
> The only possible problem I see with the on demand generation is
> that certain features like rivers would need to have the terrain
> fully instantiated if you want to calculate them based on
> gradient, and possibly erosion modeling.
This is my take on it as well, but the theory is that you make that
pass in batch and store it as an explicit database. As the display
machine needs information about the terrain in a given area, it can
generate it on the fly and then apply exceptions as appropriate.
The set of waterways and such would represent exceptions. Chopped
down trees are exceptions. Planted trees are an exception. And so
on. I have no idea of the practicality of such an approach, but I'm
considering getting back to writing such a waterflow analyzer for
the terrain generator that I have. I find this sort of thing
absolutely fascinating. Obviously, the list of exceptions can get
rather massive, but there should also be a tendency for the
exceptions to be localized and to be overwritten - lots of activity
over and over again in the same areas.
MUD-Dev mailing list
MUD-Dev at kanga.nu
More information about the MUD-Dev