When designing game engine GUI systems, I know a lot of game engines (notably Unity and Unreal) use a sort of "canvas" system. However, in an isometric game where the camera is already orthographic, this seems like it might be overkill. Everything is already dictated in terms of pixels, as is the case in an isometric game; I feel that there would be little need for such a "canvas", which seems to be the bridge between world space coordinates and screen space coordinates.
With the orthographic view, world space is more seamlessly joined with screen space, so wouldn't it be simpler to get the position of the camera, define a set of points (i.e. top-left, top-right, center, etc.) on a plane that is directly in front of the camera and render controls directly in front of the camera? Would there be any downsides to this method, or am I missing the point entirely?