This narrowness is also present in more functional uses (to which Lasswell gave equal footing) of the policy process paradigm. For example, Brewer and deLeon (and, by implication, Lasswell) never proposed that the policy process comprised a theoretic model as ascribed by Sabatier, for they certainly realized that it was not suitable to formal hypothesis testing or prediction with much precision. Rather, they viewed the policy process as a device (a heuristic, as it were) to help dissaggregate an otherwise seamless web of public policy transactions, as was too regularly depicted in political science. They proposed that each segment and transition were distinguished by differentiated actions and purposes. For instance, policy estimation was primarily an analytic activity pursued by (usually) staff analysts within an agency; on the other hand, implementation was performed by an entirely different set of actors, generally acting outside the agency, having to interact with a defined set of external clients, and occasionally having to alter literally the policy purposes as a matter of local necessity (Groggin et al 1990)