Eniac program

[jonpaul]

Programming ENIAC 1943...1945..1955
Plug boards and jumper cables..

https://lnkd.in/eNFREQuC

a first use was by John VonNeumann for  atomic  and thermonuclear bomb  physics Simulation

From an optimist in the nuclear age

Jon

[radiogeek381]

Great Post! (And it is nice to see a posting in Vintage Computers that is about a machine that predates the PC.)

A nit, I think:

I suspect von Neuman did not develop the first application. There is strong evidence that he was introduced to Eckert and Mauchly in the fall of 1944, a year before the first "real" code ran on the machine. But, according to Scott McCartney's book "ENIAC" (page 103) the first modeling code was written by Nicholas Metropolis and Stanley Frankel late in 1945.  McCartney's account looks credible,  especially in light of the personalities involved. By the fall of 1945, von Neumann was likely too involved with the discussions around the EDVAC follow-on to spend much time on coding up an application for ENIAC.

Goldstine's account "The Computer, from Pascal to von Neumann" in chapter 10 supports McCartney's account. Goldstine was there at the time, and in fact likely managed the relationship between the U Penn team and Metropolis. The book is -- for the most part -- a tribute to von Neumann, attributing much to von Neumann in a frequently worshipful tone. (I found the book very unpleasant reading.) This all suggests that Goldstine's account is reliable, as he would have credited von Neumann profusely if JvN had any part in the problem preparation.

In cross-checking I find that McCartney's account and Goldstine's account were independent. Goldstine relied on his personal correspondence from the time, while McCartney interviewed participants and drew on a recorded interview of Metropolis that was done in 1987. McCartney's book is quite well researched, and a good read.

matt

[jonpaul]

Hi Matt: Museums...

CHM:USA Sil Valley  https://computerhistory.org/
TNMOC near BP in UK:
https://www.tnmoc.org/

Mechanical computers of Charles Babbage, circa 1880s....

https://en.wikipedia.org/wiki/Difference_engine

https://www.computerhistory.org/babbage/engines/#:~:text=Charles%20Babbage%20(1791-1871),the%20method%20of%20finite%20differences.

Babbage Analytical Engine (design but never built, had CPU, storage, bus ...a complet computer programmable in gears and cams!)
https://en.wikipedia.org/wiki/Analytical_Engine

Enjoy!

Jon
Jean-Paul (EE 1968, the Internet Dinosaur)

[Sherlock Holmes]

My first real programmer job was at a computer center that was owned by a consortium of London burroughs, they all chipped in and shared the costs of a big IBM mainframe installation, manned 24/7. Anyway, some of the older guys there (I was 25) spoke fondly of the initial setup of the consortium, that was not based on IBM but on LEO. They even showed me old printouts of "LEO Code", I wish to God I had kept some of that stuff, true history.

LEO 1 was the world's first commercial business computer system, made and sold by - wait for it - a cake/tea house chain!



Before my time, must have seemed like science fiction back then.

[KE5FX]

There's a great two-part video from the mid-1990s here:



00:00 - Opening and Copyright
01:16 - Introduction with Gordon Bell
02:16 - Overview of computing in the 1930's
04:10 - George Stibitz and the BTL Mark 1
10:15 - The beginning of World War II
11:30 - Konrad Zuse and the Z3 (including an appearance by Zuse himself, which I'd never seen)
20:51 - John V Atanasoff and the ABC
29:10 - Howard Aiken, Grace Hopper, and the IBM Harvard Mark I
42:07 - Herb Grosch and the SSEC
50:35 - Closing with Gordon Bell



00:00 - Opening and Copyright
01:22 - Introduction with Gordon Bell
02:16 - Eckert, Mauchly, and ENIAC
13:59 - Moving on From ENIAC and the creation of EDVAC
25:39 - Maurice Wilkes and EDSAC
31:43 - Lyons Corporation, LEO, and the beginning of commercial computing
37:51 - Parallel computing, the IAS machine, and its offspring
50:35 - UNIVAC and the birth of popular computing
53:02 - Closing with Gordon Bell

Plenty of ENIAC content, but the section on the bakery's LEO installation blew my mind.  That thing was expensive -- how could it possibly have been justified as a business expense?  Especially at that point in time, when the effects of WWII were still being felt? 

[radiogeek381]

According to Simon Lavington's "Early British Computers" (Digital Press (speaking of historic names) 1980) The LEO design was derived from EDSAC experience and discussions with Maurice Wilkes. 

The purpose mentioned in the book was "automating office procedures such as payroll processing..."

This makes sense, and fits with a chronicle from Raymond Thompson (included in the book) indicating that a payroll program was run in January of 1953. 

BEGIN CONJECTURE:

It is possible, and perhaps even likely, that the machine was also used in inventory management or in reconciling accounts and invoices. This latter task was a huge part of the target market for mid 1950's IBM machines. For many companies inventory and invoices needed to be reconciled by the end of each work day. The process involved something like "index cards" with what we would now see as database transactions (add X to account Y, debit Z from account W). Lyons managed tea shops all over the UK. Given that some of the inventory was subject to spoilage, maintaining inventory and monitoring demand in each shop must have been as difficult as it was valuable. 

This notion is supported, perhaps, by the last entry in the chronicle for November 1954 (page 73 in the book) 

Teashops General Report for Wembley teashops has entry: "The head staff at this shop would like to give thanks for LEO. This is a wonderful time saver, and work saver, and we are grateful for it."

Perhaps they were talking exclusively about payroll, but I think it goes beyond that. (Though historical speculation is ...)

Now that I've dug a little deeper this page https://warwick.ac.uk/services/library/mrc/archives_online/digital/leo/story suggests that at least one application involved the valuation of bakery production. 



END CONJECTURE;

In any case, Lyons was big business. LEO eventually merged with English Electric, and its market extended well beyond the Lyons company. 

[KE5FX]

You should watch part 2 of the video above.  It goes into a lot of detail about LEO's role in the bakery... all except how they managed to pay for it.

[jonpaul]

Programmed by patch boards, by staff of women.

For a really vintage computer see the 1880s Babbage Engine, reconstructed in 1991 by Dr Doron Swade at NMSI in London

Enjoy


Jon

[radiogeek381]

: KE5FX

The video is on my list.

As to how Lyons could justify the cost: From my reading/research/browsing/google it appears

Lyons was the Mac Donald's of its day. It had 200 + shops all over the UK. They sold more than tea -- they sold tea-time. Baked goods, confections, ... and they served in the shop. One history claims that they were the largest caterer in UK and the continent. Their flagship restaurant could serve thousands of customers per day. (One resource quotes 500k but that is plainly rediculous, as the shop was 5 stories high and would require 200 tables per floor and a turn-around time of one minute per customer continuously over an 8 hour day.  Not gonna happen.)

The post-war shortages in fuel, vehicles, and raw materials would have made inventory management a matter of survival. Combined with rising demand and the growth of the business they likely compared the cost of development with the cost of spoilage, losses, and opportuinity lost when a shop ran out.

The US saw a similar move to automation, but here it was driven by the post-war demand and the impossibility of scaling to meet the need. The demand was sufficient (The US had lots of "big" companies) that it supported a market where IBM and the BUNCH ate the NRE. Lyons was a little ahead of the demand, and the UK could not support much commercial development on spec. The Manchester machines were along this line, and Lyons was likely to build on the government supported NRE.