Zoom and rotate the objects instead of the camera

I think this was suggested before.

When I zoom or rotate the camera through zoom and angle, what I get is a resized or rotated camera. Why not perhaps multiply the zoom with the position and the size of the objects, and rotate the position and the angle of the objects before drawing them to the camera?

So my current work around for the camera zoom is resize the camera based on its zoom but the problem is when I zoom in the camera doesn't follow well and when I zoom out it consumes a lot of RAM memory and drops the FPS on Flash with a low zoom although it's not a problem on C++.

Wouldn't it be awesome if we have that kind of camera?

Start by reviewing the camera behavior exposed through zoom and angle, along with how object positions, sizes, and angles are handled before drawing. Compare the proposed object-transformation approach with the current camera resizing workaround, including the reported Flash memory and FPS impact. Done would require an agreed design and implementation that provides the requested zoom and rotation behavior without those drawbacks.