Hi,
I also don't find it too obvious how the extention to 3d geometries is going to work, yet.
There are several cases: 1) Some concept/algorithm pairs are already fully dimension-agnostic, like point/distance: you can ask the distance between two 2D points, two 3D points, two 15D points if you want. 2) Some others are only handled in 2D. For them, the appropriate algorithms will have to be extended with 3D versions (or agnostic versions) of the algorithm. 3) Some 3D-related concepts don't even exist yet (fundamental ones, like planes, or structured ones, like meshes). They will have to be created (the *concepts*, not only one of their model) and appropriate algorithms can then be implemented. So to make things clear, when we say "GGL can be later extended for 3D and other stuff", we're not stating that everything will magically fit into case 2). Rather that we can rely on case 3) when needed to add virtually a infinity of other geometry stuff without breaking consistency. Regards Bruno