Modifier Layering System

One caveat modifiers have is they deliver a state of the geometry and this state gets used by other geometries and operations as a whole.

My wish is that modfifiers could be separated in LAYERS, so another entity could recognize the geometry up to some point and not after have been modified by ALL modifiers. That way we could get rid of e.g. dependency cycles, mesh bind oriented modifiers, etc. Take this example: If an object A is bound to the mesh of another geometry B and this geometry A gets subdivided, the binding gets lost, because object A changes polygon amount. this layering system would be a way around that.

So, the way it would work is: as we create a modifier to an object it would automatically be put to 'layer 0', and all the following modifiers also would be put to this 'layer 0'. But we would have a button to create a layer, which would then be called by default 'layer 1', and continue to put modifiers into that, and so forth. We make object B to be bound to mesh A with e.g. 'Surface Deform', then we would create a subdivision modifier to object A, so the 'Surface Deform' would no longer work because object A have changed polygon amount. So the idea is we create the 'Subdivide Surface' modifier for object A in 'layer 1' and we make the 'Surface Deform'' from object B to grab the state that the object A was up to 'layer 0' (basically discarding every layer ahead). This way we would avoid this and a lot of other problems such as dependency cycles and etc.

One other example is: a cloth which is surface deformed by a mesh and compresses it the same time.