Refactor the camera-related code

[ForNeVeR]

For now, we have a lot of copy-pasted fragments like this:

            let struct (windowWidth, windowHeight) = window.WindowSizePx
            let struct (renderTargetWidth, renderTargetHeight) = window.RenderTargetSize

            let cameraTargetX = ((windowWidth |> float32) - (renderTargetWidth |> float32) * camera.zoom) / -2f / camera.zoom
            let cameraTargetY = ((windowHeight |> float32) - (renderTargetHeight |> float32) * camera.zoom) / -2f / camera.zoom

            camera.target <- Vector2(cameraTargetX, cameraTargetY)
            camera.zoom <- min ((windowHeight |> float32) / (renderTargetHeight |> float32))
                                    ((windowWidth |> float32) / (renderTargetWidth |> float32))

We should extract that to a reusable function.

[ForNeVeR]

Closed by #151.