retroelectro
The BCC-500 was designed to get power from a motor generator so that brown-outs from the utility company would not affect it. Today, you would just use a $50 “battery backup” from the Digikey catalog.
and like many other new small businesses, BCC was feeling it. Early on, the company was able to raise two million dollars from Data Processing Financial & General Corporation (DPF&G), but that would soon come to an end. In April, Pirtle obtained a $1M investment from the University of California, Berkeley’s investment portfolio. This investment was met with the slightest controversy. Recently, the University made some poor decisions and came under scrutiny for misuse of funds. At this same time, DPF&G was in an antitrust lawsuit with International Business Machines (IBM). DPG&G invested money in developing the BCC-500 to avoid purchasing more IBM equipment in the future. When the lawsuit had settled and the BCC-500 wasn’t completed towards the end of 1970, DPF&G These are the same effects of the 1970 recession that affected Hans Camenzind and Signetics in Volume 5 of ‘We Get Technical’. Check it out.
computers. This system was attractive to SDS (Scientific Data Systems) because they made minicomputers and had been thinking of getting into the timesharing business.” When Project GENIE formally ended in 1968, the team split up in various ways mainly staying within academia and research, with some going to companies like Xerox, and a few took the route of entrepreneurship. Dr. Melvin Pirtle, Dr. Butler Lampson, Chuck Thacker, and Alan Kay wasted little time and started their own company, the Berkeley Computer Corporation (BCC). Building on the successes of Project GENIE, BCC was formed with the goal of designing the BCC- 500, a computer meant to push the boundaries of time-share computing even further.
BCC was founded on December 19, 1968, and remained in business until January 1971. Its primary purposes were manufacturing and marketing large-scale time-sharing computer systems and related equipment and offering time- sharing services on BCC-owned systems. The original staff consisted of fourteen senior- level programmers, eight senior-level engineers, and many supporting technicians, draftsmen, and programmers. Pirtle, the first Principal Investigator at Project GENIE, was the company’s President, Lampson was the head system designer, and Chuck Thacker was the engineering project leader. An internal document named ‘Supervisor’s Manual’ is available (and listed in the Suggested Reading section) that lists all the personnel with the company in October 1969. Some remarks on a large new time-sharing system Prior to going out of business, in what could have been a last- ditch effort to sell the BCC-500 prototype and stay in business, Lampson authored a report titled ‘Some Remarks on a Large New Time-Sharing System’. This document is the earliest comprehensive overview of the BCC-500.
In this report, he lists BCC’s four primary characteristics they had in mind when designing this system: ■ Efficiency, obtained through specialized hardware ■ Reliability, which depends on redundancy, error-checking, and good protection mechanisms ■ Modularity in both hardware and software ■ High-level Language Programming Here, we find the claim that the BCC-500 could support up to five hundred concurrent time-sharing users. He also discussed the machine’s unusual configuration. It consisted of two central processors, several small processors, and a rotating magnetic memory system, which allowed them to swap 250 users in and out in one second. While this sounds like it would be impressive for the time, he continues to say: ‘The Model 500 is not an unusually fast machine when measured in instructions per second, nor does it use extremely fast logic. It is the specialization of comparatively modest amounts of hardware for particular purposes, which make its overall efficiency high’, arguing that even though many of the machine’s specifications were worse compared to the
competition, it is an application- specific design, comparable or better in efficiency.’
made writing software for the system easier and quicker. The firmware was designed so that a user would get between 500ms and 750ms of allocation every five seconds. The 750ms limit was there so that the user wouldn’t notice a wildly different quality of service outside of peak usage time.
The BCC-500 would have shipped with a working FORTRAN IV and eBASIC
compiler and an SDS/XDS-940 emulator. Its operating system was a highly machine-oriented program with more than 10,000 discrete instructions, which contained fewer than one hundred discrete machine instructions. This would have
1970 – financial difficulties The recession of 1970 struck,
Figure 3: Help Wanted Ad (November 16, 1969)
Figure 4: BCC 500 Block Diagram (Some Remarks on a Large New Time-Sharing System)
we get technical
34
35
Powered by FlippingBook