August 2005

The Dream of a Lifetime

Continued from page 3

By Bill Joy

Other Dreamers
Dormouse tells the important story of what the Bay Area did for computing. But as I read the book, I found myself thinking about other early history, stories not centered on the West Coast. While the PC was born in California, its conception required important contributions from other parts of the country.

Today, PCs are highly networked, run multiple applications at the same time (much as the time-sharing computers of the 1960s and 1970s supported multiple users), and have virtual memory to support large applications. These and many other key technical capabilities originated not in the counterculture of the West Coast, but in the great universities and research labs on the East Coast, in England, and even in the upper Midwest, where I grew up.

Around the time of Engelbart's NLS presentation, a practical implementation of a different set of groundbreaking computing concepts, far beyond a mere demonstration, appeared in the form of the Michigan Terminal System (MTS) operating system.

MTS was written for a mainframe -- the IBM 360/67 -- that was one of the first computers to have virtual memory. IBM had 300 programmers writing a new operating system for this computer, but they were far behind schedule. So the staff at Michigan wrote MTS, which featured time-sharing, support for virtual memory, file sharing with protection, and many other functions in new combinations that were eventually to become key parts of the PC.

By 1967, MTS was up and running on the newly arrived 360/67, supporting 30 to 40 simultaneous users. Fully a year before MTS was finished, in 1966, Michigan began a related project, the Merit network, which would provide a way to network multiple systems. Like the early ARPAnet, Merit used minicomputers -- Digital Equipment Corporation's PDP-11s--to connect larger machines to each other.

By the time I arrived as an undergraduate at the University of Michigan in 1971, MTS and Merit were successful and stable systems. By that point, a multiprocessor system running MTS could support a hundred simultaneous interactive users, as well as remote graphics applications on computers such as the DEC 8/338 and 9/339 -- pioneering minicomputers with interactive vector graphics displays. MTS served as a campuswide network for these machines, and Merit soon connected the computers of the University of Michigan with those at other universities.

Similarly powerful systems were built on Digital Equipment PDP-10s at MIT, Stanford (SAIL), and Carnegie Mellon University, often, like Engelbart's NLS, with support from federal research funds. Markoff recounts in passing what I had forgotten (if I ever knew it) -- that Steve Jobs and Steve Wozniak were hanging out at SAIL long before the famous Jobs visit to PARC. SAIL, and similar systems, had much greater importance in the birth of the PC than is generally acknowledged. In my view, these systems underpin, as much as Engelbart's work does, personal computing.