Just over 35 years ago, Carl Hewitt and his graduate students published a model for computation based on concurrent message-passing Actors. Now the demands of many-core computers and cloud-based software are thrusting that model to the forefront. In this conversation with host Jon Udell, Hewitt explores hardware-enforced cloud privacy, paraconsistent logic, and scalable semantic integration. This terminology is still unfamiliar to many--coming from the frontiers of research. But the discussion boils down to a few basic principles that apply universally to carbon/silicon-based life forms including ORGS (Organizations of Restricted Generality). First, a carbon/silicon actor can change only its own behavior. To influence others actors have to send the right messages and then work it out. Second, there will always be pervasive inconsistency in our information systems. But actors can negotiate effectively even in the presence of inconsistency and rapidly recover when things go wrong.
Carl E. Hewitt is Associate Professor Emeritus in the Electrica Engineering and Computer Science department at the Massachusetts Institute of Technology. Hewitt is known for his design of Planner, which was the first programming language based on procedural plans that were invoked using pattern-directed invocation from assertions and goals. He is also known for his work on the Actor model of concurrent computation, which influenced the development of the Scheme programming language and theπ calculus, and served as an inspiration for several other programming languages.
This free podcast is from our Jon Udell's Interviews with Innovators series.