The authors focus on the subset of multi-channel signal processing systems that detect and analyze events in long-term signals. Although ECG analysis programs are typical of this class of systems, signal processing programs from many other problem areas exhibit similar behavior. The authors' basic premise is that a considerable degree of uniformity can be identified in systems of this sort. The goal of the research is to construct a general-purpose framework that exploits this common structure--a programming environment within which the programmer need only address aspects of an analysis problem that are specific to the application. A description is given of the design and performance of SEEK/PE, a system which extends SEEK/MC by addressing apsects of multi-channel signal processing other than those related to learning and pattern recognition.