A key intuition underlying our work is that simple hierarchies are not rich enough to capture complex structures. In fact, we believe that any single ontological structure is insufficient. As a result, we have been exploring a variety of mechanisms that make it possible to view implement a system from multiple perspectives. One thing we believe about these multiple perspectives is that in order to have significant power they must be able to crosscut each other. That is, one perspective must be able to organize the implementation in fundamentally different ways than other perspectives.

The thread that has been common to our projects is rooted in the fact that the normal rules of modularity do not suffice for describing quality software.