A Letter from Prof. Corbató (10/30/00)
It is hard to believe that the last instance of Multics hardware is about to be extinguished. But of course the influence of Multics has probably achieved immortality through the admirable evolution of Unix and more recently Linux.
I have often wondered why we so stubbornly worked so hard to make the system survive. My own take on it is that we were young and wanted to make a dent in the psyche of the industry which in those bad old days was incredibly shortsighted. And I think we did. With the creation and persistence of working systems, Multics became a paradigm for a comprehensive solution to a host of system problems that even today are not fully addressed in many systems. In no particular order, some of the key ideas I think of are: a hierarchal file system, system backup policies, rings and memory protection, symmetric multiprocessing, paging and memory management, dynamic linking, access control, and a full character set.
I can think of two reasons Multics survived so long. One was its malleability and evolvability which I attribute to the use of a higher level language (despite PL/1!), the major effort to organize and maintain the system with functional modularity, and the avoidance of cute or obscure names in modules. And the second reason was the idealistic zeal of all the Multicians who worked so hard on a system they believed in.
And I think we also owe a special debt of gratitude to the original triumvirate of Bob Fano, Ed David and John Weil who had the guts to start it all, to John Couleur and Ted Glaser for their courage in messing with the hardware when the system was nothing but silly putty, and to Elliott Organick who told the world what we were up to.
With fond memories,