and I found that it was very easy to overlay some buttons in a corner. And I'm definitely one of the less practiced graphics programmers around here...
IMO the keyword here is "easy" - only two specific LMV functions were used into my code above for example.
If someone wants to draw into the MapViewer canvas, he must do it through the viewer drawing engine, which is technically a different way of drawing than LCL. Even more so that all objects there have a map coordinates, no notion for another (screen) coordinate space, thus anchoring an artifact on screen means constant movements into the map space. It requires some effort, that's why I wanted to know the benefit, according to some other person, of embedding such extra gadgets.