it would be better if texture blitting as well as 3D objects has tree-like hierachy which pass translations including move, scale, rotation and show/hide states. so everything can be managed in groups.

i found texture tile to be not precisely repeating:
i have a plane created by

plane=ExtendedPrimitives.createPlane(4096,20); ExtendedPrimitives.transformUVCoords(plane,0,0,256,256,0,0);
tex.setClamping(0); plane.setTexture(tex); plane.compileAndStrip(); plane.build();

texture image size is 256x256.
say i look at position (0,0,0) of the plane, i see some part of the texture. then i translate the plane to (4096,0,0), in theory i should see the same texture, but i saw the texture changed. no matter what parameter i put into the transformUVCoords call. i don't know why.

that works , thanks.
but beware, this also affects texture blitting. i added an empty world and set the nearclip to 1, in order to reset the matrix for 2D blitting.

Which renderer are you using for that?

lwjgl hardware opengl renderer

i am trying to do multipass rendering of a single world, dividing a world into near/middle/far parts. render far parts before rendering near parts,  like this:

world.setClippingPlanes(250.0 , 20000.0); world.renderScene(framebuffer); world.draw(framebuffer); framebuffer.clearZBufferOnly(); world.setClippingPlanes(5.0 , 250.0); world.renderScene(framebuffer); world.draw(framebuffer); framebuffer.clearZBufferOnly();
world.setClippingPlanes(0.1 , 5.0); world.renderScene(framebuffer); world.draw(framebuffer)

but the end result is not what i expect. is my method wrong? how can i achieve the multipass effect i described?

i get 0.5 and 1.5 for minfov and maxfov, those are about 30 and 75 degree in horizontal direction. but i want larger range.

And honestly, I don't think that your GPU doesn't support per pixel fogging. That's a feature almost as old as GPUs. What makes you think that it doesn't support it?

say my camera is placed at (0,0,0) and there is a large quad of size 20000x20000 at (0,10,0), with each vertex at the same distance to camera (at 45,135,-45,-135 degree). fog ends at 5000.
when the quad is rendered, its color is totally fog color. because all vertices have the color of fog, they make all pixels have the same color. this is the behavior of per-vertex fog.
if the 2 vertices which were behind the camera are moved forward so that they are a little bit in front of the camera, then the fog effect looks correct.
right now i am using Intel Atom Z3740D with onboard GPU.
anyway i'll try to use shader to render the quad, i did it once with other engine. i was just curious how software renderer would do in this case.

ok. i have never done wiki before.