Taming large complex information systems

Empirical evidence from entity relationship analyses of corporate information systems shows there is explosive growth in the logical complexity needed to specify the behavior of the totally detailed integrated whole. It is not surprising therefore that the history of integrating large corporate information systems is a tale of woe. How can the complexity be mastered? The empirical evidence also shows that information systems have a cellular data structure defined by loops over relationships between entities. These loops almost always "commute" in the mathematical sense used in category theory. they can be used to define a cellular partitioning of the system. This leads to the general conjecture that an effective model for large stable systems is a cellular structure with a selective message passing medium between cell boundaries. Such a model restrains local context from interacting across the system to the point of intractable complexity. What is selected as local context and global context is then the essence of the system.