July 2012 archive

Jul 25 2012

Book review: A computer called LEO by Georgina Ferry

AComputerCalledLEOThis is a review of “A Computer called LEO” by Georgina Ferry, recounting the story of the first computer developed for business use by J. Lyons & Co, the teashop and catering company.

Lyons formed in 1884, a spin-off from a family tobacconist company whose traveling salesman realised that there were few reliable teashops around the country, furthermore catering at large events such as the Great Exhibition was poor. Over the next 30 years or so the business grew, with a chain of teashops, and smarter establishments such as the Corner Houses and Trocadero. The teashops were supplied by Lyons own manufacturing and delivery service.

By the 1930s Lyons had approximately 30,000 workers, as such it was one of Britain’s larger employers. 300 clerks were used to tot up daily takings on mechanical calculators. Clerical work had risen in important during the second half of the 19th century with numbers rising from 70,000 in Britain in 1851 to 2 million in 1901. The company had a department of Systems Research led by a Cambridge mathematician, John Simmons, who the company had recruited in 1923, the hiring of such a graduate was a novelty at the time. The Systems Research department was interested in the efficient running of the business.

By this time various items of office machinery were commonplace, things such as filing cabinets, typewriters, mechanical calculators, and punch card readers. Telephone exchanges were in place, the electronic valve had been invented in the early years of the 20th century and magnetic storage devices were starting to become available. By the 1930s people such Oliver Standingford in Lyon’s Stock Department were talking about machines which would combine these elements, although he was not clear on the detail of how this would be done.

The Second World War then intervened, Lyons cut table service from its teashops as labour went short. Various people gained useful experience in electrical engineering through the wartime developments in radar, and possibly codebreaking. We now know that Colossus, a computer used for code breaking, was built at Bletchley Park during the war but it did not become public knowledge until 1974. In the US ENIAC had been developed at the Moore School in Philadelphia to do artillery range calculations. This was not a secret and immediately after the war, Oliver Standingford and Raymond Thompson visited from Lyons; they had a broad brief to investigate American business methods but it was ENIAC which really captivated them. Fortunately, their US trip put them in touch with more local expertise in the form of Douglas Hartree at Cambridge University who was building a computer, EDSAC, for the Mathematical Laboratory.

Lyons decided fairly quickly to construct their own computer, which was to be based on the EDSAC machine; US machines such as they were could not be purchased because of currency restrictions and there were no computer manufacturers in the UK. From the start LEO I (the first computer) was different, Simmons saw the computer fitting into a system of “scientific management” and as such LEO was crafted to exactly fit the role he foresaw for it based on detailed knowledge of the company’s processes. In some senses computing for business was more demanding than the computation done in the Mathematical Laboratory and other scientific laboratories: business computing had large demands for input and output (imagine a payroll system – it needs to read in details of each employee and print out the results), it had lower tolerance for failure (payroll failing to run has a serious impact on employees) and calculations could be more “complex” than mathematical ones in the sense that more steps in calculation and more conditionality was required. It was at Lyons that the art of flowcharting was developed. The first live duty that LEO carried out was in 1951, it was made public in 1955. It’s interesting to note that Charles Babbage had highlighted the potential for automation in both manufacturing and mathematical operations in his book “On the Economy of Machinery and Manufacturers”, published in 1832.

There were to be two further LEO computers, developed by a separate company, Leo Computers Ltd however things did not go well. The computers themselves were technically advanced, and the Leo Computers method of going into a business and closely examining their processes before writing programs and delivering a system combining both hardware and software usually had excellent results. However, this had the unfortunate side-effect of losing their best staff to their clients. Other problems were afoot: Leo Computers Ltd although nominally a separate company was under-resourced both financially and in personnel with development engineers also acting as salesman. The parent company, Lyons was struggling – victim of a family business mentality which put increasingly useless family members at the heads of divisions.

In 1964 Leo Computers Ltd was merged with English Electric, with Lyons divesting itself of any responsibility, following this union the LEO line died although the final computers in the series were installed by the Post Office, and continued to run there, in places, until 1981.

In contrast in the 1960s IBM were able to make an investment of $5billion on their System 360 computers – a compatible range designed to fit every need. They had a ready market in the US both of businesses willing to buy, unlike their British counterparts, and a government who bought locally first. Faced with this opposition, the British computer industry struggled to compete.

Focusing on the LEO computer makes this a human scale story with central cast of characters, but it also provides a wider view of the field in the years after the Second World War. The book makes clear how J. Lyons & Co had a system of management, and personnel in place which were ripe for computerisation; the developments in the 1930s made it clear that electronic computers were in the air. Large scale failures of computer systems in both public and private sectors are onging, John Simmons was rather insightful in his intimate coupling between business process and software system.

References

1. My Evernotes are here

2. The web page of the Leo Computer Society is http://www.leo-computers.org.uk/

Jul 02 2012

Book Review: The Geek Manifesto by Mark Henderson

The-Geek-ManifestoThis review is of Mark Henderson’s book “The Geek Manifesto: Why Science Matters“. It starts with Simon Singh’s lengthy libel battle with the British Chiropractic Association, which sets the scene for the rest of the book.

In some senses this is a book about “my people”, many of the events described were ones I watched unfold online. For example, I wrote back here, about David Nutt’s sacking from the Advisory Council on the Misuse of Drugs. I remember the Boots 10:23 homeopathy protests in which geeks overdosed on homeopathic remedies, I even got ever so slightly involved in Science is Vital. The acknowledgements go out to many people I follow on twitter, and the links to other books I have read, “Adapt” by Tim Harford and “Bad Science” by Ben Goldacre, are explicit. I’d also add “Freakonomics” by  Steven D. Levitt and Stephen J. Dubner to this genre.

After the opening Simon Singh case there are chapters on the role of science in politics, government, the media, the economy, education, court, healthcare, and environmentalism. Random fact I picked up: only two scientists have appeared on the BBC Question Time programme in the last 8 years – somewhat shocking given the wide roster of the great and the good who appear on the programme. I liked the practical examples of reducing crime in Cardiff, led by Professor Jonathan Sheperd, a maxillofacial surgeon frustrated and distressed by the number of young people admitted through Accident and Emergency with severe cuts from pub brawls. He collected detailed evidence of incidents which he passed on to police, later publishing research on the outcome of this approach by comparison with a similar city without such intervention.

I also liked the stories from the University of Southampton on fibre optic research and Shankar Balasubramanian (at Cambridge University), whose gene sequencing company, Solexa was recently acquired for $600million. Too often scientists make the public assertion that spending money on research benefits the economy, but then say that you can’t possibly direct them into particular areas of research – the old blue skies argument on which I blogged here and here. These case studies are convincing on the benefits of longer term, university-based research.

The section on climate change and environmentalism fits somewhat with my opinions – in principle I am green but find the official face of the environmental movement (Greenpeace, Friends of the Earth, the Green Party, The Soil Association) difficult because they go beyond the environment to an anti-capitalist world view in which good science is optional; particularly in campaigns against genetically modified crops and nuclear power. However I don’t believe the Greens can be blamed for the, typically, right-wing opposition to the consensus on anthropogenic  climate change. I wrote a bit about climate change here. This also touches a little on the Sense About Science organisation, one mentioned several times through the book, they are no doubt useful but I have my suspicions about some of their positions, particularly on climate change – this article in The Skeptic captures some of that unease.

The book is very well described as a manifesto: it’s aimed at the geeks of the title. It picks up on a theme of increased levels of activism amongst scientists, a growing skeptic movement and the newly formed comedy / science intersection (Dara O’Brien, Robin Ince and so forth). I’m a scientist, I’ve been a scientist for 20 or so years, and I’m undeniably geeky – I write programs for fun, yet I find the “geek” label uncomfortable.

There has been a twitter campaign to send this book to every single MP, all I can say is I hope they don’t read the second chapter. Henderson, a historian by training, seems rather overwhelmed by the forensic powers of scientists and dismisses the idea that anyone else might have them. I think his attitude is summarised in his reference to introducing more scientists to parliament which will apparently

“…rapidly raise the game of the non-scientists around them”

I’m a scientist, and I find this embarrassingly patronising. His remedy for bad science in the media is rather more subtle and nuanced than his approach to politicians. I think in the end, in politics, he highlights some of the important things – the need to engage with MPs rather than leave it to others; the need to engage early when positions are not set in stone and the need to recognise that MPs face other pressures: a recent Early Day Motion in support of homeopathy is a case in point: many MPs signed because they believed, that many of their constituents were passionately in favour of it, the support of Prince Charles probably helped them with their decision too.

The Behavioural Insights Unit in the Cabinet has a recent paper “Test, Learn, Adapt: Developing Public Policy with Randomised Controlled Trials” by Laura Haynes, Owain Service, Ben Goldacre and David Torgerson (pdf) which is exactly the approach to government that Henderson is promoting. It perhaps highlights that a frontal assault on MPs is not the way in: the civil service is important. A second strand, which I would propose, is casting the whole experimenting thing into a business context. One of my own political preferences is for governments that do things competently (regardless of political tendency), in business this might be labelled something like “Excellent Execution”. Doing things competently isn’t a particular skill of scientists: it’s something everyone can aspire to, and scientists by no means have a monopoly on the routes to competence, or even universally exhibit themselves.

Once again, I am tricked by Kindle – the text of this book ends at about 60%, it is very heavily referenced!