Proactive computing

Reviewer: John S. Edwards

I could hardly wait to read this paper, since I am always seeking to learn exactly what the word “proactive” means relative to the word “active.” In many cases, the real meaning has more to do with the notion of being in favor of action than it does with action itself. Inside the Beltway (near where I live), it often means “make a commotion about something and maybe nobody will notice there is precious little actual action.” Here, Tennenhouse seems to refer to a future world in which computing elements outnumber humans by several orders of magnitude. He refers to the predictive work published by J. C. R. Licklider in 1960. Tennenhouse tries to look ahead to the next long period of computer evolution. He identifies an agenda to reexamine: the boundaries between the physical and virtual worlds, the time constraints at which computation is applied, and the movement from human-centered to human-supervised computing. He states that “focusing significantly greater attention on them will enable a new mode of operation” and calls this new mode “proactive computing.” I cannot hazard a guess why. Nonetheless, he defines three research areas for proactive computing: physical coupling, faster responses, and removing humans from the loop and placing them above the loop. He focuses on embedded computers in future devices as ways to optimize control structures. A figure shows today's computer science focusing on interactive computers, and tomorrow's on embedded computers. I am struck by an example from long ago, when the boss of a newly minted electrical engineer asked him to construct a turn signal for a car. While the complex of relays, capacitors, and other components he created could indeed function as a turn signal, it was nowhere near the elegant solution invented for the purpose. Take this simple example and make it several orders of magnitude more complex. The author makes a plea for computer science to address these problems. Left alone, new graduates will not have the tools necessary to use the implements economically. The author's clarion call for the reinvention of computer science should be given close attention (despite his use of that awful non-word).