pages 17 and 18:
Hackers, Wild Ducks, and Skunk Works: Timesharing and Virtualization from 1959 to 1968pages 13 and 14:
By Sean P. McBride, 30 March 2013
For Itm 554: Operating System Virtualization, Jeremy Hajek
Illinois Institute Of Technology
While Norm Rasmussen and Bob Creasy began planning their timesharing project,
another part of IBM discovered a possible reprieve from the MIT fiasco. The University of
Michigan had long been a flagship IBM campus, second only to the MIT Computation Center.
After MIT’s snubbing of IBM for Project MAC, Professors Bernard Galler and Bruce Arden
sensed an opportunity to overcome IBM’s traditional predilection for elite East Coast universities
and position the University of Michigan as IBM’s new flagship campus (Akera 2008).
Contacting IBM and the National Science Foundation, these professors suggested that they could
build a System/360 timesharing system, as long as IBM followed their specifications for
hardware changes to enable dynamic addressing. IBM tentatively accepted to this proposal, but
soon changed course after reaching out to customers to help the professors gather a list of
requirements. Because customer demand for timesharing far exceeded their expectations, IBM
decided to instead develop the timesharing system internally as an official supported product
released alongside a new System/360 model with DAT capabilities. However, the University of
Michigan and five other clients under non-disclosure agreements would form the “Inner Six” and
help IBM steer the direction of the product via SHARE user group committee (Varian 1997).
In August 1966, the TSS development team informed the ‘Inner Six’ that it was unlikely
that the Model 67 would be released before December or that TSS would be released before
April 1967. Even worse, the first release of TSS would only be for “experimental,
developmental, or instructional use” (Pugh, Johnson and Palmer 1991, 362). The reasons for this
were myriad. According to a member of the TSS architecture group, “OS/360 hadn't settled town
sufficiently when TSS began, and there was too much of a rush to completion” (Goodwin 2009).
Out of the ‘Inner Six,’ the University of Michigan was especially frustrated by these delays, as
they had promised that that timesharing services would be available by the fall semester of 1966.
Professor Bernard Galler accused IBM of “attempt[ing] from the beginning to build a system
and included everything except the kitchen sink” (Akera 2008). Having obediently cast aside
their aspirations to develop a timesharing system in-house to support TSS two years earlier, the
Michigan Computer Center now believed that they had misplaced their faith in the abilities of
IBM development, leading them to resume development of their own Michigan Terminal System
for the System/360.17 These adverse announcements rocked the company similar to the loss of
Project MAC, leading IBM CEO Tom Watson Jr. to admit in the 1966 Annual Report that IBM
had “experienced delays in meeting our original objectives for… time-sharing systems” (Pugh,
Johnson and Palmer 1991, 362).
17 This system became the Michigan Terminal System, which was operational six months after delivery of their
Model 67, fully developed by late 1968, and run across a consortium of several universities by 1969. (Akera 2008).
Akera, Atsushi. "The Life and Work of Bernard A. Galler (1928 - 2006) ." IEEE Annals of the
History of Computing (IEEE Computer Society) 30 (2008): 4-14.
Pugh, Emerson W., Lyle R. Johnson, and John H. Palmer. IBM's 360 and Early 370 Systems.
Cambridge: MIT Press, 1991.
Varian, Melinda. VM and the VM Community: Past, Present, and Future. Office of Computing
and Information Technology, Princeton University, SHARE Inc., 1997.