|Gordon Bell And DEC - The Mini Computer Era|
|Written by Historian|
|Tuesday, 19 August 2014|
Page 2 of 2
While Bell was away DEC carried on building machines - but nothing that was as successful as Bell's machines. The PDP-10 was another attempt at the niche that the PDP-6 was supposed to fill - i.e. a big multiprocessor machine. It still didn't sell in huge numbers and almost sank without trace.
Even though Bell was doing his best to be an academic he remained a consultant to DEC and played a large role in the design of the PDP-11.
This was arguably the first modern mini-computer. It had multiple general purpose registers, a wide range of addressing modes, a stack pointer and stack operations. It was also modular in the sense that it used a standard bus - the unibus - which allowed easy connection of peripherals.
The unibus was the first example of a bus that was almost more important than the machine it served! The PDP-11 was immensely popular and gave DEC the lead in the minicomputer market in the 70s.
In 1972 Bell decided to take a sabbatical to write a book on digital design but Ken Olsen asked him to be vice president of engineering at DEC. He accepted and decided to attempt to put some of the ideas he had been working on at Carnegie-Mellon concerning multi-processor machines and Large Scale Integration (LSI).
In 1975 DEC announced the VAX-11, the Virtual Address eXtension of the PDP-11.
The VAX was the first of the super-minis and it was another success. It turned up in applications that had previously needed mainframes and in many ways it is best regarded as a small, low cost mainframe - but with one important difference - networking was built into its over all design.
A company could install a range of VAX machines and connect them together in a networked (Ethernet) time sharing system.
In 1978 Bell thought of starting a company to produce personal computers - but didn't because the VAX was taking up so much of his time. Eventually though he did tire of the big company. The VAX system had a life of its own and there was little left for him to do.
In 1983 he had a heart attack and after recovering decided not to go back to DEC. He considered various academic options but decided to set up Encore Computers to build multiprocessor super machines based on his multimax architecture. In 1986 he also set up Dana Computers to build single user supercomputers and he became the assistant director of the National Science Foundation's Computing Directorate where he set up a number of national supercomputer centres.
Looking to the Future
All of this doesn't reveal another side of Gordon Bell.
As well as writing a number of well known books on digital design, he wrote papers that predicted the future of computing - a multiprocessor one of course - and placed public bets with colleagues about when the multiprocessor future would arrive on every desk top.
In 1972 he formulated Bell's Law of of Computer Classes, which describes how types of computing systems (referred to as computer classes) form, evolve and may eventually die out.
The definition of the law states:
Roughly every decade a new, lower priced computer class forms based on a new programming platform, network, and interface resulting in new usage and the establishment of a new industry.
Bell considered his law to be partially a corollary to Moore's law which states "the number of transistors per chip doubles every 24 months" but the driver of Bell's law was the availability lower cost components.
It postulates that technology advances in semiconductors, storage, interfaces and networks enable a new computer class (platform) to form about every decade to serve a new need. At the time the law was proposed classes included mainframes (1960s) and minicomputers (1970s) and in an interview in 1992, by which time networked workstations and personal computers were firmly established Bell said in an interview:
"Twenty-five years from now...computers will be exactly like telephones. They are probably going to be communicating all the time"
To promote the development of high-performance computing applications, the Gordon Bell Prize was established in 1987 for the application of parallel processing to real scientific and engineering problems. Funded by Gordon Bell, it is administered by the ACM and IEEE and is presented annually to recognize performance achievements that are innovative and generalizable. The 1994 winners of the prize were so successful, with performance figures in the TeraFlop range for the time, that Bell doubled the prize because he was so pleased!
Curating the past and Lifelogging
Together with his wife Gwen and with the assistance of DEC, Bell founded the Digital Computer Museum in 1979 and was a founding board member of its successor, the Computer History Museum. In 2003 he was made a Fellow of the museum "for his key role in the minicomputer revolution, and for contributions as a computer architect and entrepreneur."
Gordon Bell joined Microsoft in 1995 but had already been involved in the establishment of Microsoft Research in an advisory capacity. His first focus was telepresense, technologies that allow a person to engage with others from a remote place or time - summed up by Bell as being there without really being there, then.
Currently he is the "experimental subject" of a Microsoft Research project inspired by Vannevar Bush's hypothetical Memex computer system. MyLifeBits is an experiment in life-logging, the practice of wearing computers in order to capture continuous data from lifetime experiences which can be be accessed with speed and ease. The project and is an attempt to fulfill Vannevar Bush's vision of an automated store of the documents, pictures (including those taken automatically), and sounds an individual has experienced in his lifetime. For this, Bell has digitized all documents he has read or produced, CDs, emails, and so on.
He continues to do so, gathering web pages browsed, phone and instant messaging conversations more or less automatically using wearable devices and has co-authored a book in connection with this project. It is available in two editions Total Recall: How the E-memory Revolution Will Change Everything (2009) and Your Life, Uploaded: The Digital Way to Better Memory, Health, and Productivity (2010).
Man or Machine
You may think that these later achievements stand a good chance of having a longer term effect on the computing world than just having invented a few really good machines.
I would argue that the influence of the PDP series of machines and the VAX goes deeper and is much longer lasting. The PDP series proved that computers had a value in a smaller environment. They proved to many people that a smaller cheaper and more personal compute was needed. The PDP 8 even looks like the archetypal first home computer the Altair, complete with flashing lights and keyswitches.
They also formed the breeding ground for the people who would one day push the microcomputer revolution forward. The PDP 11 in particular, with its regular modern architecture, demonstrated what good computer principles were all about. I'm sure that C and Unix were inspired by its internal design and there must be lots of other lesser known software.
There also must be lots of programmers who still look back on the PDP 11's design as a golden age compared with the ugliness of the current generation of the x86 family.
or email your comment to: email@example.com
|Last Updated ( Tuesday, 19 August 2014 )|