The nature and assumptions of information processing models vary in an apparently haphazard fashion within the disciplines that utilize them. As a result, while each individual model or theory may be more-or-less internally consistent, it is often difficult for students of information processing to evaluate and compare these models. The present paper has two goals. First, we present an embryonic attempt at construction a taxonomy, or "model space," in which a large number of information processing models may be represented as different locations in the space. Dimensions include system architecture, nature of the primary independent variable, the nature of the state space (or "activation;" e.g., discrete or continuous), whether or not the system has memory, and whether or not the system is stochastic. Second, we illustrate one useful consequence of constructing such a taxonomy by showing how a rigorous methodology developed for one class of models can be generalized and applied to another class within the overall taxonomy. To this end, we show that systems-factorial technology, which allows for architecture identification in discrete flow systems, may be generalized for use as a diagnostic tool in continuous flow systems as well.