Modifying the vertex position within the vertex shader is a straightforward way to offload some computation from the CPU to the GPU. It also eliminates the possible need to transfer vertex buffers between the GPU memory and main memory in order to modify the positions.
The main disadvantage is that the updated positions are not available on the CPU side. For example, they might be needed for additional processing (such as collision detection). However, there are a number of ways to provide this data back to the CPU. One technique might be the clever use of FBOs to receive the updated positions from the fragment shader. In a later recipe, we'll look at another technique that makes use of a newer OpenGL feature called transform ...
