From Cakes to Apples
It all started with ENIAC and EDSAC. No, that’s wrong, it started with Alan Turing and the ‘bombe’. In fact, that’s misleading, too. At the very least we should go back to Charles Babbage and Ada Lovelace. This is a familiar problem: when you are seeking to explain some important innovation, you want to find the crucial first step that initiated change, yet innovation is seldom like that. It’s like claiming printing began with Johannes Gutenberg around 1450 AD. He did invent the first movable type printing system in Europe, but printing from woodblocks developed many hundreds of years earlier, while movable type printing appeared in China in the 11th Century (generally ascribed to Bi Sheng in 1040 AD!). I’d better begin again.
It might have started with the abacus, back in the 3rd Millennium BC, during the Babylonian era. An abacus is a calculating tool. They are still used in markets in some areas of Asia, the Middle East and Africa. However, a decisive leap forward came with the invention of the mechanical calculator in the 17th Century, and, with various improvements, these were still widely used in the first half of the 20th Century. I still have the one my father relied on. The next step, in the 19th Century, came from Charles Babbage working on various powered calculators, of which the most important (although unfinished) was the ‘analytical engine’, a general purpose calculating device, which was intended to operate with both data and instructions (a program), to be input using cards (based on the Jacquard loom punched card which controlled weaving operations).
Charles Babbage is one of my heroes, together with Ada Lovelace, the only legitimate daughter of Lord Byron, who is claimed to have been the first to recognise that Babbage’s machine had applications beyond pure calculation. A determined self-taught mathematician, she was in constant correspondence with Babbage, publishing the first mathematical process she had devised and intended to be carried out on his machine. She is widely regarded as one of the first computer programmers. [i] Years later, Alan Turing’s work on the bombe during the Second World War drew on Babbage’s ideas, now developing them into an electromechanical device.
With that background, we can now return to ENIAC, the Electronic Numerical Integrator and Computer, generally regarded as the first general purpose electronic computer. It had been developed for the USA Army to calculate artillery firing tables, although it was also used assessing the feasibility of an atomic bomb. It began operations in 1945, was formally handed over the US army in 1946, and remained in use until 1955.
Before I go on, one interesting side note: ENIACS’s first six primary programmers were all women. The skills needed in those days comprised knowing how to input ENIAC programs, but also understanding ENIAC’s inner workings. The team was drawn from a larger group of women employed as ‘computers’ who processed mathematical formulas on mechanical calculators, generating the results required for a scientific study, or an engineering project. A group of brilliant young women, they were seen as secondary to the men building the hardware, and were unrecognised for most of their lifetimes. However, some fifty years later the six were inducted into the Women in Technology International Hall of Fame in 1997. [ii]
Over in the UK, a similar project was under way, beginning in 1947 at Cambridge University. The result, EDSAC (Electronic Delay Storage Automatic Computer), became operational in 1949, the second example of an electronic stored-program electronic computer to go into operation. It remained in service until 1958. In turn, EDSAC begat LEO, a computer I was to meet in the 1960’s, but to get there we need a little more background.
Older UK residents will remember Lyons. J. Lyons & Co. was a family-run business, founded in the late 19th century to provide catering for major events such as the annual Buckingham Palace garden parties. [iii] It grew rapidly, taking on many other activities including bakeries, ice cream, hotels, tea blending and packing, and the well-known teashops and Corner Houses. [iv] They were well-known, and as a teenager I would pop into a Lyons teashop for a cup of tea and an Eccles cake. Sadly, the last Corner House, at Marble Arch, closed in 1974, and the teashops in 1981.
Back in the late 1940s, Lyons kept abreast of new technologies they could use in their business. They sent a team of management high-flyers to the USA to see what could be learned from the practices developed during the war years. They arrived as the first electronic computer, ENIAC, was being finished, and saw how it could help in a variety of company operations. Based on the visit, Lyons decided to acquire a computer, not easy when there just a handful in development! Having heard about Cambridge University’s work on EDSAC, the Lyons Board accepted a senior management proposal they should work with the university, provide some funding for development, and, in return, Lyons would be allowed to copy some of EDSAC’s design to build its own computer, later to be named the LEO or Lyons Electronic Office. In November 1951, LEO was operational, first carrying out financial reviews of bakeries, a business application involving considerable data but little by way of complex calculations. However, it was processing speed that mattered, and more applications followed for different Lyons product divisions. Lyons established LEO Computers Limited, which was, for a while, the pre-eminent British computer company, manufacturing commercial computers and processing data for a range of clients, including leading companies ranging from Ford to Kodak, as well as for Government departments, municipalities and the Post Office. Quite a change from its catering and food businesses! Farsighted, the next stage was to move beyond manufacturing computers to being a general service provider, and this was what caught my attention in 1966.
That year, I was a postgraduate student in need of a research topic, and, unlike a single Mr. Bingley at Netherfield Park, I was not surrounded by a wide range of possibilities. Trained as a social anthropologist, I might have considered some suitable, largely unknown tribe in the New Guinea Highlands or the Upper Amazon, but married, with two young children, that didn’t seem a sensible option. What kinds of ‘primitive society’ could I find closer to home? Then I read about LEO, and it’s innovative Computer Service Bureau. Struck by the idea I could study this new area of business as it were a distinctive community, I had a topic. Unfortunately, the Department of Social Anthropology did not consider it a suitable one, and I ended up in an obscure part of the University’s Department of Engineering.
Research often proceeds in a far from predictable fashion. Initially wanting to study the relationships, myths, rituals and dating practices of a computer bureau (alright, not the dating practices), I quickly found it was the computer programmers who were the most intriguing feature of English Electric Leo Marconi. This was a very different world from ENIAC’s early days. Now programmers were seen as critical to computer use, and programming, writing instructions in machine code, and seen as important as computer hardware engineering. However, by this time the field was full of men, jostling aside talented ‘computer women’.
A second development was the programming task was differentiating. At one level there were those programmers who translated mathematical instructions into machine instructions. When writing operational programs for the computer itself, this task was undertaken in what was a somewhat esoteric machine code. For other tasks, working on specific applications, there were computer languages, one of which was FORTRAN. Before long, I was writing some programs myself, to aid in my research data analysis. I was lucky: in 1961 FORTRAN IV had been released, thankfully a version which removed the necessity to learn and use machine code.
Sitting ‘above’ programmers were another group, systems analysts. Their task was to look at an application, planning delivery routes for example, and then work out the mathematical analysis that would be needed to implement this application. As I was soon to discover, while some aspects of programming and systems analysis required mathematical skills, logic and an understanding of systems was pre-eminent. To my surprise, several of the people I met in the bureau were classicists, a group, I was told, with excellent logical and analytical skills. In a quickly emerging hierarchy, above them were senior systems analysts, who would talk to a bureau customer, understand what it was the customer wanted to know, and then provide an overview of what the programming task would have to achieve.
If that all sounds rather prosaic, it wasn’t. While I was supposed to be collecting data, I spent quite a lot of time gossiping with analysts and programmers, especially on the limits of what could be done. As an example of this, which I heard about at an early stage, was in relation to an application English Electric wanted. The company built electrical supply generators, huge devices, central to which was a central spinning core, often weighing several tons. There was a problem, I was told. Every time they were to build a new and bigger generator, they would take the existing design, and simply increase everything. That would work, but only up to a point, then suddenly the next iteration would fail, and the rotating core would disintegrate (a very dangerous moment). Once this happened, the engineers would go back to the drawing board. Surely, I asked, this was a computational problem? In the 1960s, that wasn’t the issue. Rather there wasn’t an adequate mathematics to describe the behaviour of rapidly revolving, complex and very heavy objects. A program wasn’t possible, and so it was left to engineers to ‘guess’ what might get around the latest disaster. I suspect that challenge no longer exists, but back then, I was fascinated. I got caught up in discussions about rules of thumb, approximations, and ‘fudge factors’. The engineers became my focus, a group of scientists who were always living on the edge of what science could offer them, and where they had to rely on tacit knowledge and guesswork. Somewhere along the line, I became determined to study these engineers, just before the new CEO at English Electric was briefed about the businesses he’d inherited, and learned there was a research student on the loose. I was booted out, and my research collapsed.
Despite being unable to complete my study, my involvement did have other consequences. As I mentioned, I began writing programs. Cambridge’s computer, Titan, was far from titanic, but unlike the Atlas computer, on which it was based, this device had real, cached memory, using a ferrite magnetic core. It initially had 28K of memory, but this was expanded first to 64K and later to 128K. To put that in perspective, that is considerably less than the memory in a basic digital watch (an iWatch has far more, reserving some 6 GB for its own use, in addition to space for user data storage!). Everything else Titan relied on was virtual memory, on disc drives and magnetic tape. As I am typing, my desktop computer has a 500 GB internal memory storage, roughly 5 million times more than Titan! Indeed, “by the standards of today’s technology LEO could be regarded as belonging to the stone-age. Today a hearing aid or mobile smart phone has many thousand times the storage capacity and speed of LEO and whereas LEO occupied a large room, a hearing aid can be sufficiently small to be hidden in the ear”. [v]
My research gave me an unexpected opportunity, to attend the closing down of the original LEO computer. Computers in those days took up a lot of space, a medium sized room, and you could stand inside many of the working parts. They used mercury delay lines, some of which made audible sounds, giving rise to the slang term “mumble tub” for these devices: they were noisy. [vi] For this last day, a programmer put together a sequence of cards to set off various sounds. A series of dark-suited executives stood, tears running down many faces, listening to LEO playing God Save the Queen. Equally remarkable, as soon as LEO was switched off those same senior people pulled out screwdrivers and pliers and were busily removing parts as souvenirs.
Moving forward by fifty years, it is hard to imagine that time. My use of Titan was to run some correlations on data collected from staff in a hospital. I would leave a long box of punch cards, mainly data, at the computer centre and wait. Quite often I would get my programming slightly wrong, and received nothing back. At other times I would find myself bumped off the queue, and had to wait another day. Little did I know, but most of the computer time was used on such tasks as analysing inverse Fourier Transforms on data from the One-mile Radio Telescope! [vii] I didn’t even know what a Fourier Transform was, let alone its inverse, and I still don’t.
In 2011, I took the fateful step of abandoning Microsoft and IBM clones, and became an Apple user. We have Apples everywhere: Macs as desktops, iPads, iPhones and a MacBook. Without my having to think about it, data and materials flow seamlessly from one device to another, but also without my noticing, Apple has me firmly in its grasp. Quite unable to do much about it, Apple controls my applications; I had to abandon Skype on my old iPad as it was no longer supported on that ancient (seven years old!) version of an operating system. To use Skype when travelling I had to buy a newer iPad! Now I can Zoom! My desktop daytime workhorse is amazing, powerful, fast, with many applications I never need to use. The upside is I can do just about anything I want to do. The downside? Unlike fifty years ago when I could and did write programs, I am effectively an ignorant and fairly unskilled user. From the lost world of Lyons cake shops and LEO to the empire of Apple, I guess I’ve become happy but relatively powerless, captured by IOS. No matter, it must be time for my afternoon cuppa and an Eccles cake.
[i] I am extremely fond of Babbage and Lovelace. Maboth Mosely’s 1964 book Irascible Genius (Hutchinson) gives a good overview of the development of the Analytical Engine and Ada Lovelace’s role in the process.
[ii] Janelle Brown, Women Proto-programmers Get Their Just Reward, Wired, 8 May, 1997
[iii] See T Harding, Legacy: One Family, a Cup of Tea and the Company That Took On the World, Heinemann, 2020. Only recently did I rediscover their factory was close to my childhood home!
[iv] This review of Lyons and LEO is in https://warwick.ac.uk/services/library/mrc/archives_online/digital/leo/story/
[v] Op cit
[vi] http://www.rfcafe.com/references/popular-electronics/electronic-mind-remembers-popular-electronics-august-1956.htm
[vii] As revealed in https://en.wikipedia.org/wiki/Titan_(1963_computer)
