Records the state of all the vertex buffer bindings required to provide a vertex declaration with the input data it needs for the vertex elements.
Why do we have this binding list rather than just have VertexElement referring to the vertex buffers direct? Well, in the underlying APIs, binding the vertex buffers to an index (or 'stream') is the way that vertex data is linked, so this structure better reflects the realities of that. In addition, by separating the vertex declaration from the list of vertex buffer bindings, it becomes possible to reuse bindings between declarations and vice versa, giving opportunities to reduce the state changes required to perform rendering.