Talk:LEO (computer)

Main memory cycle time
Are you sure about 1500ms this is 1.5 seconds oughtn't this to be at least micro seconds? Rjstott


 * I just thought of the same; 1.5 sec seems a little strange to me too. I tried to find some info on the net, but failed, so I'll check with the LEO society.  I certainly wonder where the info in this article came from.  --Wernher 18:03, 5 Dec 2003 (UTC)

As I recall the main memory cycle time of LEO III was 13 micro-seconds. I later worked on some re-furbished LEO IIs and one of the few things I remember about frequency is the carrier of the mercury delay lines being 12.6 Mhz. I am not (yet) remembering how to relate this to access time (other than word length) or instruction execution time, but I think there were 256 words of 39 bits in each mercury "tube". I am not remembering anything about start/stop/sync bits at this time either. —Preceding unsigned comment added by 72.93.163.236 (talk) 18:18, 7 January 2010 (UTC)

Colossus connection

 * J. Lyons and Co., one of the UK's leading catering and food manufacturing companies in the first half of the 20th century, sent two of its senior managers to the USA in 1947 to look at new business methods developed during the Second World War. During their visit they came across digital computers then used exclusively for engineering and mathematical computations.

Although this story is widely published, what is not so well known is that the chairman of Lyons worked for MI5 or MI6 during the WWII and that he was well aware that computers could be of use in civilian business. The two gentlemen probably did go to the states, but given the absolute secrecy that surrounded Ultra until the 1970s, it would have made a good cover story to explain why a tea shop wanted to build a computer. Professor Frank Land was one of the early programmers on LEO and when I studied under him he gave some lectures on programming LEO and an introduction to the computer's history. I seem to remember him mentioning the MI5/6 connection but not the trip to the states. Does anyone have a written source to confirm that my recollections are correct or even the name of the post war chairman of Lyons --PBS 18:06, 6 August 2005 (UTC)


 * The chairman at the time (1950–1956) was Major Montague I[sidore] Gluckstein (J. Lyons and Co.):
 * Major Montague Gluckstein, chairman of J. Lyons and Co. Ltd., The National Archives
 * p. 37 ''Reflections on the History of Computing: Preserving Memories and Sharing Stories by Arthur Tatnall (ed.) "Brief biolgraphical sketches of a number of Lyons and LEO people can be found in Leo, the First Business Computer; P. Bird, Hasler Publishing, 1994, pages 200–212, The following people in alphabetical order – are noted: ... Montague Gluckstein ..."
 * -- PBS (talk) 11:37, 7 July 2015 (UTC)

LEO III or LEO II
The sentence "The first LEO III was completed in 1961." in section 3 seems erroneous to me. Shouldn't that be LEO II, as the LEO III was built after the takeover by English Electric a few years later (according to the information in the same paragraph)?

M. Schouppe —Preceding unsigned comment added by Special:Contributions/ (talk)

The first LEO III was installed in 1962. --Aspro 19:04, 31 August 2007 (UTC)

In January 1962 I attended a Leo III course at Hartree House. The first LEO III was "under wraps" above our class room, there was an operational II there as well. The merger with EE was in 1963, I transferred and moved to Kidsgrove in October of '63. The LEO III was designed, developed and built WELL BEFORE any influences from EE - most of which were negative. —Preceding unsigned comment added by 72.93.163.236 (talk) 18:10, 7 January 2010 (UTC)

I contributed the above paragraph, I had not INTENTIONALLY failed to sign it, I did not and still do not know how to sign it. Reg Burgess - LEO Jan '62 - Apr '65 -- I suppose this entry will appear as "unsigned" as well :D
 * Hi Reg, you can sign using --~ . That will automatically insert --, your Wikiname (or IP address if you are not logged in), date and time. There is also a button for that in the editor, just to the right of the Bold and Italics buttons; it looks like a pencil signing a signature. --Jhertel (talk) 20:17, 3 February 2015 (UTC)

Re: The LEO speakers
Some operators managed to get the computer to play tunes. Listen to example here: But I am not sure that this is strictly encyclopaedic.--Aspro 19:04, 31 August 2007 (UTC)

I certainly got chain printers to play tunes on DEC printers.

Anyway what I wanted to say is that "LEO: The first businesss computer" by Peter J Bird (privately printed I believe) might be worh adding. My mum worked on LEO on the fifties (as a minion of course-- you didn't let women near the machine) 92.40.235.169 (talk) 00:46, 2 October 2008 (UTC)trewy@live.co.uk (Si Trew)

This is anecdotal - probably no use for Wikipedia, but may be of interest. In the 1950s my uncle worked for the Gas Board (later British Gas) and was one of several people from their accounts department sent to see an early Leo demonstrated in their offices at Whiteleys Department Store, with a view to eventually computerising their own accounts. Even then the Gas Board would have been a customer worth several hundred thousand pounds a year to whoever supplied their computers. When they got there the guy in charge basically said "This is the Leo computer, it's working out our accounts" and went off, giving them no explanation as to what it was doing, what all the pretty lights were for, how errors etc. were handled, etc. etc. They would have learned more by reading a brochure. So they waited an hour, the guy in charge never came back, and eventually they left with no idea whether it would be in any way suitable for their needs. Eventually they bought their computers somewhere else... Marcus Rowland (talk) 12:09, 6 January 2011 (UTC)

I worked for EEM which became ICSL then BARIC. I led a shift of Operators working on a LEO III among others for 4 years or so. Yes, you could tell what the machine was doing based upon the sounds the computer made via the speaker. Basically this was reverse engineering though, in that as the machine was performing know operations, it made a specific sound. For example, in later years when tape drives were added ( 1968) a tape read error and retry had a set sound and so on. Commands were given initially via paper tape, and thus one became proficient in reading and writing this media.. and thus one could give the machine instructions that could emulate a tune. — Preceding unsigned comment added by 71.228.150.166 (talk) 01:52, 23 June 2012 (UTC)

I worked as an operator using ICL 2900's, which I suppose were a descendant of LEO, and which also had a speaker attached to the CPU, which was useful diagnostically. Loops (a sort of whining) and hangs (silence) were particularly distinctive. Most of the time it was like radion static but various program in the order sales suite had keynote sounds. The sales rep analysis sounded quite different. — Preceding unsigned comment added by 62.3.255.103 (talk) 22:17, 27 March 2013 (UTC)

Designed by Oliver Standingford and Raymond Thompson?
Not borne out by the text. Surely more likely designed by Hartree, Wilkes and Pinkerton, project proposed and managed by Standingford and Thompson?Moletrouser (talk) 10:17, 24 August 2010 (UTC)


 * Yes, I think you are right. According to Hally, LEO I was modelled closely on EDSAC, which was designed by Hartree and Wilkes, and the construction was managed by Pinkerton who reported to Thompson. Standingford and Thompson obviously had a key role in initiating the construction of LEO I, but probably had little involvement in its technical design. I have changed "designed" to "overseen" in the lead. Gandalf61 (talk) 11:58, 24 August 2010 (UTC)

The mighty LEO 326
I was one of the programmers on the Freemans (SW9) LEO 326 in the 1960s, I pass on some comments for you.

The LEO 326 had a 48 bit word 4 bits of which stored the program identity and controlled multi-programming the rest of the word was divided into two half words each with a sign bit and parity bit. Effectively one was working with two 20 bit half words or one 40 bit word.

It had two primary registers the A and B together with a radix register C which allowed it to do arithmetic in Pounds shillings and pence, Pounds and ounces, Feet and inches etc.

CLEO programs were prepared using 10 forms, somewhat analogous to the DIVISIONS of COBOL. I cannot remember them all but 5 was data and 10 was certainly the procedural code.

Programs were identified by a five digit number.

Only the first four characters of a data name were used by the compiler, which did not check uniqueness. Data fields could be alphanumeric (6 bit characters), hexadecimal (4 bit), Long or short integers (long binary integers were a full word and short a half word). One could also specify that a data item was monetary (in those days pounds shillings and pence).

CLEO programs were converted into INTERCODE which was then assembled into machine code. In practice INTERCODE was virtually machine code and used the machine code command set with data names rather than addresses.

The Freemans machine of the mid 1960s had 5 x 4K banks of memory, 4 Elliott 1,000 cps paper tape readers, 7 x 96K ch/sec Ampex tape decks and 2 1,000 lpm special Analex printers.

Feeding this machine was the largest data preparation shop in Europe with 240 Olivetti machines operating a single entry regime using check digit verification and check totalling to avoid the need to verify. It worked very well indeed.

The capacity of these machines was awe inspiring, for when properly programmed they were very fast indeed. I recall some benchmark work was done in the 70's with high end IBM 370 series mainframes failing to keep up with them.

Johnrwise (talk) 09:45, 20 August 2012 (UTC)

External links modified
Hello fellow Wikipedians,

I have just modified one external link on LEO (computer). Please take a moment to review my edit. If you have any questions, or need the bot to ignore the links, or the page altogether, please visit this simple FaQ for additional information. I made the following changes:
 * Added archive https://web.archive.org/web/20051123125148/http://www.jpbowen.com/publications/thes-leo.html to http://www.jpbowen.com/publications/thes-leo.html

When you have finished reviewing my changes, you may follow the instructions on the template below to fix any issues with the URLs.

Cheers.— InternetArchiveBot  (Report bug) 08:54, 14 December 2017 (UTC)

A Commons file used on this page or its Wikidata item has been nominated for deletion
The following Wikimedia Commons file used on this page or its Wikidata item has been nominated for deletion: Participate in the deletion discussion at the. —Community Tech bot (talk) 23:19, 10 February 2021 (UTC)
 * Leo I Computer.jpg