Few projects depend upon a flat design approach.
You may group your SQL objects under schema/name-spaces/users/operational tasks and even further sub-divide those actions by the database/instance those objects are on or even the servers/cluster that houses the work.
All of this could be easily shown with a user defined tree structure where you can move items up or down the tree or define new tree nodes to organize your work as it is needed.
This would help with projects with a great number of diagrams as well as re-factoring data structures (having your existing data structure on one branch of the tree while another branch holds the target structure).
It would even aid in separating physical and logical designs.