Fred Longstaff Genius Programmer

I have been fortunate to work with some amazing people in my career.  One that stands out in my mind is Fred Longstaff.  I was reminded of him recently when I had lunch with an old friend who talked about his design and implementation of a real time executive.  I met Fred on a project where we were using a mini-computer to control a nuclear power station.  Fred created a very efficient low overhead executive that ran high priority programs in memory and lower priority ones were swapped in and out.  It was an amazingly simple design but worked like a charm.  My friend took the design and used it in his real time systems for the last thirty years.  It was such an amazing design.

I was then involved in developing systems for the Bruce Nuclear Power Station.  The computer system we were using had an very poor development system.  The editor and the assembler were incompatible and the filing system treated the disk like a magnetic tape.  I convince my client to bring Fred in to build a development system.  Fred estimated it would take him forty days.  He had to prepare the programs by using coding sheets which would be key punched by two key punch operators.  It was the most amazing thing to watch.  Fred sat with his feet up on the desk thinking for about a week, scribbling a few notes.  The following week he took his feet off the desk and started coding.  He coded so fast that the keypunch people could not keep up with him.  He coded for a week.  Remember he had to use the old development system to assemble and edit his programs.  He was developing a completely new disk operating system, file manager, editor, link editor, and putting a front end on the existing assembler.  Then he started assembling the programs.  It was quite incredible but at the end of the forty days we had the most amazing development system.  A team of developers used that for years to develop all the code for Bruce 1.  I expect it was also used for future stations but I do not know.  How he did all that work with old development system and then move to the new one.  I think he also built a debugging tool as well. 

It was quite an experience.  He was really a genius.  I think it is such a gift to be able to design and build things like that.  I am still in awe.  The system fitted together exactly the way we needed it for the work we had to do.

Does anybody remember Fred?

  1. Roger D Moore Reply

    What is the value of cosine zero?
    A] FP6000#2 Back in Malton we were unsure as to exactly to what the NRE was doing with the machine. Someone said “they throw bombs in the ocean”. I never saw the machine although I did write a small amount for it. :
    A1] NRE was promised some transcendental software subroutines of moderately unusual specification. I think the list included trigonometric, exp, log. The unusual feature of the trigonometric functions was that argument of sine/cosine was a number in -1.0 to 0.9999.. range representing an angle from -180• around to 180• again. Result was in the -1.0 to .999… range. Numbers were FP6000 double precision (47 bits twos complement).
    Fred Longstaff was in charge of the panic-driven task force which produced these routines in a few weeks. I was summoned from Regina for about a month as a team member. There were one or two other people (perhaps John Chapman and Jim McSherry). I was first assigned to create some infrastructure in the form of Fortran callable subroutines to do arithmetic with large numbers.
    Fred directed me to a monograph by Clenshaw on calculating with Chebyshev polynomials. (I cannot find exact reference with Google but it was published by HMSO around 1955). After writing the infrastructure routines, I was assigned sine/cosine by Fred. (Other team members were coding other functions). When I had completed the routine for sine/cosine I had a dispute with Fred bout the value of cosine zero. I suggested -1.0 with overflow but Fred vetoed this and so the result became the largest representable number.

    A2] I never knew that #2 had exotic devices such as a track-ball and plotter connected.

  2. Jim Reply

    Fred died in the 90’s I believe.
    Fred was part of a team at Ferranti Packard that designed the only Canadian computer call the FP 6000 in the early 60’s. Fred wrote the Fortran Compiler in the early days of Fortran. The whole team split up when Ferranti ran out of money and sold the design to a British computer company. Fred and others formed ESE, a ground breaking communications company, doing digital filtering before its time. Another group of seven fellows formed IP Sharp, another neat company doing consulting and creative things with computers.
    I knew several of these people and they all were quite creative and unusual group.
    Anybody else remember any stories?

  3. Tom Reply

    Tom wrote:
    I will be interested to know if anyone else remembers Fred. I really only knew him by reputation. I am sure that Ian has fond memories. Do you know if Fred is still around?

  4. Ian Reply

    I don’t remember too much about Fred. I do remember the story on the
    blog, but because you, and probably others, told it to me back in the
    very early 70’s, but I remember it the context of the Gentilly nuclear
    power station just outside Trois Rivieres. I spent quite a bit of time
    working on various pieces of software and documentation for the Gentilly
    station, including inheriting support for the real-time executive that
    Fred had written. I met Fred only a few times, as I recall. Whom did he
    work for? I just tried finding out more about him by searching on
    Google. Was he with Ferranti-Packard involved in the FP-6000?

Leave a Reply

captcha *