Harvest SceneGraph for useful Editor feature code

[gorlak]

Right now Editor is heavily dependent on the SceneGraph object model, but moving forward we probably want to make massive changes due to #21. This will probably significantly change how complex functionality in the Editor work (like undo/redo user changes, undo/redo object/component create/delete, selection, picking, and manipulation).

[gorlak]

The elephant in the room for myself, @antiuniverse and @aclysma. Do we have a game-plan for this? I have tons of knowledge on the items listed... so feel free to shoot questions to the google group and I can share.

[aclysma]

My opinion is that we need to focus on getting the "game" and the editor minimally functional as soon as possible. So I suggest that we essentially prototype the editor integration independently of the scene graph. Ignore selection, undo, etc. Just get all the entities in a scene to list, be able to click on something in the list, modify the properties, and hit save. Then, once we get that far we can look at options to port the scene graph into the newly prototyped design. If the prototype/scene graph turn out to be incompatible, we won't be heavily invested and we can scrap bits of the prototype for something else, but benefit from the understanding gained from prototyping something integrating in the editor.

[gorlak]

I think a good way to move forward with this is to completely tear down the SceneGraph::Node classes and move useful code into library functions that we can consume via brand-new Editor-only components. I am pretty confident that there is more code to delete than there is to save, but there are some great bits in there we want to preserve.