[SIGCIS-Members] whirlwind, radar and real-time tracking

Fri Apr 23 19:39:21 PDT 2021

The thing to remember is that 1951 is very early in the history of modern computing and there are not that many functional EDVAC-style computers around. Add to that the non-obviousness of using one for real-time control and the thing that needs explaining is that there was one at all, not that there weren’t more.

The first EDVAC style computers weren’t applied to production work until 1949, with the well-documented cases of the Manchester Mark 1 and EDVAC in the UK. SEAC in 1950 was the first well-documented case in the US, as it’s not clear whether BINAC (1949) was ever applied to real problems. By 1951 there were several others, including LEO and the Pilot ACE in the UK, the IAS computer apparently partially operational, EDVAC itself in partial operation at BRL alongside ENIAC, but few enough that it wouldn’t be that much work to produce a complete list. Most of the projects were held up by the difficulty of producing reliable storage devices, more than anything else. 1951 is also the year of the very first deliveries of computers marketed as standard products.

This history highlights the exceptional place of MIT in the emerging Cold War order, as people like Forrester and Everett took advantage of connections and capabilities built up during MIT’s wartime defense work to mobilize vast amounts of money towards their project, justified by the need for new defensive capabilities to guard against the USSR’s newly developed atomic bombs. As well as Atsushi’s book, the obvious references points here are Stuart Leslie’s book _The Cold War and American Science_ (actually about MIT and Stanford) and Paul Edwards’ _The Closed World_ which has a great chapter on SAGE.

Not at all coincidentally, the unmatched federal largess that Whirlwind and other projects brought to MIT and its Lincoln Lab over the course of the 1950s explains why it was pretty much unique in the late-1950s and early-1960s in having a handful of computers lying around without constant queues of official work to be done on them, which in turn created the potential for self-selected undergraduates and even non-students drawn to its potential to develop MIT’s famous hacker culture.

It’s not just Whirlwind. All of the examples that spring to mind from the 1950s and early-1960s of using computers for realtime control, or of making them more rugged and compact, come from Cold War applications such as mobile control centers, missile guidance, the manned space program, bombers, nuclear reactors, etc. All this is explored in chapter 4, “The Computer Becomes a Real-Time Control System” in the forthcoming Haigh & Ceruzzi _A New History of Modern Computing_ which begins with Whirlwind and finishes with the Space Shuttle, along the way exploring the origin of important technologies such as minicomputers and chips as byproducts of the work done for these Cold War projects.

Best wishes,

Tom

From: Members <members-bounces at lists.sigcis.org> On Behalf Of Hansen Hsu
Sent: Friday, April 23, 2021 8:15 PM
To: Guy Fedorkow <guy.fedorkow at gmail.com>
Cc: members <members at sigcis.org>
Subject: Re: [SIGCIS-Members] whirlwind, radar and real-time tracking

Hi Guy,

If there were any digital computers used for real-time applications before Whirlwind, I certainly am not aware of them.

>From my reading of Atsushi Akera’s Calculating a Natural World, Jay Forrester’s decision to build a digital von-Neumann architecture computer for real-time applications, initially for aircraft simulation, was actually rather non-obvious. It would have been much more straightforward (and cheaper) to build an analog device. But once the Navy backed out, the general purpose nature of the Whirlwind allowed Forrester to repurpose the machine for new patrons, the Air Force thus SAGE.

So as far as I know, this is still a “first” claim that has held, as much as we want to be careful about using the “f” word. Again, if anybody knows differently, we’d all like to hear about it!

Best,

Hansen

On Apr 23, 2021, at 5:50 PM, Guy Fedorkow <guy.fedorkow at gmail.com <mailto:guy.fedorkow at gmail.com> > wrote:

Bonjour Pierre,
  I sometimes fear that I live inside the MIT bubble, where Whirlwind's "obvious" status of launching real time computing is a given...  Do you know of a good reference to put the work into context with the rest of the world?
  I'll admit that as I wrote the words about innovation and vacuum, I was tempted to add "but Whirlwind was so astonishingly expensive that it's hard to believe there could have been competitors who weren't as prominent".
  I can see that once the Whirlwind team had shown it could be done, and that there were a few more computers around, lots of teams would jump on the idea.
  Thanks for your advice!
/guy

On 4/23/2021 7:19 PM, Pierre Mounier-Kuhn wrote:

Hello Guy,

As you certainly know, Whirlwind is considered to be the first digital computer designed for real-time computing, particularly for radar tracking and interception guidance or assistance to tactical decision.

I have studied air-defense systems in Europe, particularly in France: The first projects involving digital computers did not appear before the mid-1950s, at IBM France and in a small Paris company, SEA, which was also developing digital control devices for machine-tools. At that time, several similar digital computing projects were being developed, in the USA of course (at GE, in the US Navy with Univac, etc.) but also in Britain and in the USSR.

It is true that "innovations like this rarely occur in a complete vacuum": The Whirlwind was built at MIT, one of the world's richest environments for innovation in electronics and defense systems, which had worked on a previous analogue calculator project for the US Navy. Air defense systems already existed, based on radars, telecom lines, control rooms and command centers: The idea to replace manual operators with a computer to process signals and make decisions faster "naturally" came to various people in the context of the Cold War. The Whirlwind was nevertheless a leap forward in technology, logical design and use.

Hoping that these simple remarks help you.
Best,
Pierre Mounier-Kuhn
CNRS & Sorbonne Université, Paris

----- Mail original -----
De: "Guy Fedorkow" <guy.fedorkow at gmail.com <mailto:guy.fedorkow at gmail.com> >
À: "members" <members at sigcis.org <mailto:members at sigcis.org> >
Envoyé: Vendredi 23 Avril 2021 22:41:03
Objet: [SIGCIS-Members] whirlwind, radar and real-time tracking

Greetings Colleagues,
   I've been working on restoring a 1951 Whirlwind program, written at
MIT, used to demonstrate real-time tracking of aircraft with radar for
the purposes of guiding an interception (the Cold War was in full flight
in the 1950's).  This work ultimately led to the massive SAGE air
defense network in the US.
   You can see some rather informal preliminary notes on the work at
https://www.historia-mollimercium.com/whirlwind/WW-Track-while-Scan-Draft-Notes-v1.pdf
   The program does work in simulation; you can see a four-minute video
of the simulator running an intercept at
https://www.historia-mollimercium.com/whirlwind/Track-while-scan-Apr-23-2021.mp4
   Spoiler alerts: The original really did display moving dots on a CRT,
but the graphics are "spartan" to say the least.  And nothing in
particular happens when the intercept actually happens.

   Would anyone know of contemporaneous work involving digital computers
for either radar tracking or real-time computing around 1951?  I think
all the familiar digital computers from those years were used in
applications where batch operation was perfectly acceptable, e.g.,
computing ballistics tables.
   Innovations like this rarely occur in a complete vacuum, but I don't
see references to any similar digital computing projects.
   If anyone has pointers, do let me know!
Thanks
Guy Fedorkow

_______________________________________________
This email is relayed from members at sigcis.org, the email discussion list of SHOT SIGCIS. Opinions expressed here are those of the member posting and are not reviewed, edited, or endorsed by SIGCIS. The list archives are at http://lists.sigcis.org/pipermail/members-sigcis.org/ and you can change your subscription options at http://lists.sigcis.org/listinfo.cgi/members-sigcis.org

_______________________________________________
This email is relayed from members at sigcis.org <http://sigcis.org> , the email discussion list of SHOT SIGCIS. Opinions expressed here are those of the member posting and are not reviewed, edited, or endorsed by SIGCIS. The list archives are at http://lists.sigcis.org/pipermail/members-sigcis.org/ and you can change your subscription options at http://lists.sigcis.org/listinfo.cgi/members-sigcis.org

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.sigcis.org/pipermail/members-sigcis.org/attachments/20210423/ada20a0a/attachment.htm>