SDI Seminar

Speaker: Gerald P. Popek,UCLA & Platinum Corp.

Date: November 12, 1998
Time: Noon
Place: Wean Hall 8220

Computer Engineering: What Research is Important


It is argued that the computer field is fundamentally unlike natural sciences; that much more than in biology, physics, or medicine, we create the phenomena we study. If the form of the system under consideration is unworkable, or ineffective, then the study of it is likely to be useless, and therefore undesirable. Typically, that rendition is discarded and another, more amenable version is created. Furthermore, only those forms which are economically viable are real candidates to meet this criterion. A variety of examples are selected from recent history to illustrate and argue this viewpoint, from computer memory systems, systems software, and the Internet.

The implications looking forward of this perspective will be explored, in terms of what is likely to matter and the effect of the uncertainty of those predictions.


Jerry Popek is Chief Technical Officer at Platinum technology, one of the dozen largest independent software companies in the world. He is also a professor of computer science at UCLA. His research is known for creating and proving the idea of "single system images," by which collections of machines without any shared memory, whether connected at high speed or widely geographically distributed, can operate and are programmed like a single transparent environment. Such software designs are commonly employed today, especially in highly resilient systems. He also ran the largest unclassified computer security research program in the U.S. Jerry founded Locus Computing Corp, at one time the largest independent developer of Unix system software. His current research interests are focused on distributed systems.