you were both right, making an array of reference to chunk does not allocate anything it just creates reference.
then I have to instanciate a class and link it to this reference.
I see no problem in line of theory. Keep in mind however that wrapping the world to reset it around the origin might be more problematic than expected. I can tell for sure Bullet (a physics API I'm using) wouldn't allow you to do so. If you're going to write your own physics, keep in mind the periodic "re-centering" operation will likely have to be supported somehow.
well I'm planning on coding everything, i'm doing this for fun^^
what kind of other type of infinite world algorythm is there out there? I searched the first 10 pages of google and each camera implementation is the same.
the camera never moves, the world does.
therefore I don't see any other way to do it.
It is possible to do something like what you have drawn in the first two images, let the camera move forward for a few chunks (say 1024*16 units) and keep streaming (letting the model "grow", without changing camera position). Then, wrap to the origin only when a certain rationale is met, such as distance from the origin > 1024*16. This is fairly easy: just subtract a constant such as 1024*16 along a certain axis.
why? basically you are asking me to do exactly the same thing but to increase the "reset" distance from 1 chunk to 1024 chunk.
what will it changes?
Edited by sliders_alpha, 30 April 2012 - 10:50 AM.