Interview with Martyn Thomas CBE

Martyn was born in Salisbury in the UK in 1948.  His mother was a nurse, and his father was a Government scientist working at Porton Down as a spectroscopist.

Martyn attended Uplands Primary School where he did the Eleven Plus, which he passed to get a place at Bishop Wordsworth’s Grammar School in Salisbury.

They had a rather strange scheme whereby, if you seemed to be doing well, you did your O Levels in four years and went straight into the sixth form, and so he was, he feels, ridiculously young when he passed his A Levels. This was a problem, because it meant that, even though he took a year between school and university, he felt “stupidly young and immature” by the time he got to university.

Martyn took maths, physics, chemistry and biology at A level as he was influenced by his parents, especially his father and he really enjoyed science. With hindsight, he wishes he had done some arts subjects to a high level, and really studied those, but, he was brought up to feel that science was supreme and that that was very definitely where he ought to be going.

Martyn hadn’t got particularly good A Levels, so he thought he’d stay on and do S Levels, essentially re-taking them at a higher level. He spent a couple of months at the school helping with a science exhibition and spent a very large amount of time playing snooker, especially with the French master.

For his gap year he worked as a junior scientist at Porton Down, in one of the labs doing biological experiments.  He was doing practical science in a secret establishment, doing classified work to a high level of science, but with hindsight, in a way that was remarkably casual. The actual science was very rigorous, and they were working with dangerous agents. He remembers that there were self-injection kits at the end of the lab just in case someone contaminated themselves with a nerve agent for example, so that they could stand some chance of recovering.

Martyn was under pressure from his father to pursue a science career.  His mother really wanted him to be a doctor, and there was no way that with his background and A Levels he was going to get into a medical school.  However, there was a potential route to becoming a doctor by doing a biochemistry degree first, and then converting to medicine.  That is what he told University College London he planned to do when he applied to go there.

He now feels he’d probably have done better academically if he had studied something completely different, maybe even one of the humanities.

Martyn studied at UCL from 1966-1969, because he was not doing well at biochemistry he decided not to be a rebel and simply did the courses that were recommended in his third year. One of these was computing and numerical analysis, taught by the professor and head of the computing centre there, Paul Samet. The requirement before you could go on the course, was that, in the very first week of January in your final year you did a, a one-week course in FORTRAN and he loved it. Martyn recalls that it was completely magical, that he could write something in a mathematically looking language and instruct a £2 million computer to do calculations for you, and the next day you would get back a listing of paper that showed what happened, and the outcome. Martyn fell in love with the feeling of power, of being able to calculate things that it would have been hugely difficult to do by hand, and to do it very quickly and succinctly.

After university Martyn worked with the University College London Computer Centre between 1969 and 1973.

After finishing his degree, it was clear to Martyn that he was not going to enjoy biochemistry, so he signed up to the Inns of Court School of Law and became a student member of the Middle Temple, with the initial idea that he might become a patent lawyer.

Unexpectedly, Paul Samet, the professor of computer science who had been teaching Martyn asked if he would be interested in a one-year job carrying out a research project for the National Computing Centre.  And of course, he jumped at the chance. After six months, Samet offered him a permanent job in the systems department in the computer centre, looking after the mainframe computer, to make sure that the operating system was running effectively.

These were the early days of mainframe computing, and, and it was a big IBM 360 65, running OS/360. And, it crashed, often, sometimes several times a day. Martyn was sufficiently interested that he went on an IBM course to learn about the internals of the operating system which he found complexly fascinating and took on the responsibility for finding the errors when the system crashed. Working with the IBM support staff, he managed to improve the reliability of the operating systems so it was down to crashing once or twice a month.

Paul Samet had made an arrangement with the University of Delft, the Technische Hogeschool in Delft, to provide them with operating systems support for writing a new ALGOL 60 compiler. Martyn spent the next couple of years, between London and Delft, learning an awful lot about ALGOL 60, about compilers, and, IBM operating systems. He recalls that it was a very interesting time.

It was obvious to Martyn he was fascinated by IT, and secondly, that he was very good at it. He was really interested in the detail and seemed to have a good grasp of the systems view of computing.  With hindsight, he attributes this to the systems view of human chemical systems and their interrelationships which he learnt from his time studying biology and biochemistry.

In 1973 Martyn moved on from UCL to work at Standard Telephones and Cables, subsidiary of the American company ITT, where he worked for two years and helped to design their first generation of computer-controlled PABXs, private telephone switches.

Whilst at STC, Martyn was introduced to a top-down graphical design method invented by Doug Ross of the US company SofTech, called structured analysis and design technique (SADT). He really liked the semi-formal, top-down, functional development style of developing software, and became interested in much more formal structured techniques such as the mathematical techniques that Dijkstra and Hoare were promoting.

From STC Martyn moved on to work for the new South West Universities Regional Computer Centre (SWURCC) on the campus of the University of Bath.  The South West Universities (Bath, Bristol, Exeter, Cardiff and UWIST) had a network of ICL System 4 Computers and had been given one of the new ICL mainframe 2980 computers to set up a regional computer centre on the network. Martyn was appointed to run the compilers and utilities development and support team alongside his university friend David Bean who was to run the operating systems team.

At the beginning of the project the most important thing to get going at the South West Universities Regional Computer Centre was FORTRAN, because that was the main language that people needed. But they were also big users of ALGOL W, a development of ALGOL 60 that Niklaus Wirth had developed. For the first year and a half they couldn’t run anything on the ICL mainframe, because, even once the mainframe had been successfully installed they really didn’t have any stable software. It was very early days.

Martyn remembers that it was a real challenge doing the development work needed. But, it was successful, and they managed to get an initial service up. But then, quite rapidly they needed to solve the problem of how they were going to replace ALGOL W. The users said that, what they wanted was ALGOL 68, as the replacement language. And so, Martyn looked around at how that might be possible.

 ICL didn’t have an ALGOL 68 compiler, so Martyn persuaded ICL to fund them to produce a real product quality ALGOL 68 compiler. ICL agreed to do this on the condition that it was written entirely to the standards of their other compilers. So, Martyn’s team had to use all the interface software, it had to be written in S3, the programming language for the system, programming language for the 2900 series, which incidentally and perhaps ironically was a derivative itself of ALGOL 68. And, it had to be provided with all the documentation, the standard manuals, the same format of error messages, all the things that made it look like an ICL standard product and part of their language suite.  And so, Martyn and his team at SWUCN took on a contract to do just that, working with the compilers group at the Royal Signals and Radar Establishment in Malvern (Philip Woodward, Sue Bond and their colleagues – who had developed an Algol 68 parser and intermediate code generator) and Oxford University Computing Centre (who also had a 2980 mainframe and had started work on a code generator)

During his 8 years at SWURCC, Martyn was appointed as Systems Manager and then Deputy Director of the Regional Computer Centre.

 He had realised that there was an opportunity to build all kinds of software that the university needed, and to do without it costing the university any money.  So, he started taking on commercial contracts on behalf of the university and hiring people on short-term contracts to do the development work. This proved to be a success but following a review of the funding of the Regional Computer Centre, Martyn was instructed to close down the commercial work that they had been doing.

Despite his efforts Martyn could not find a way to find a way to continue this work within the university, but he had made a commitment to the people he had brought to Bath that he would guarantee their jobs.

As part of the solution to Martyn’s predicament Praxis was founded in 1983 with his former colleague David Bean. The two friends had always talked about the possibility one day of setting up their own software house, and Martyn had got a team of people that needed continued work and he had some customers who needed their existing contracts completed and delivered. David Bean who had since returned to work for Logica VTS was feeling that he had come to the end of what he wanted to do, so they agreed in principle that was what they were going to do and started doing some planning for it.  Martyn recalls “And then, David, as is his wont, precipitated action by telling me he had resigned, and we were starting two months later.  And, and so off we went.”

Left with a set of contracts that they now no longer had the wherewithal to deliver against, the university were delighted to get rid of them including developing a FORTRAN compiler for ICLs Distributed Array Processor, which gave Praxis work to start up with.

ICL won a contract to do quite a lot of research and development projects for the Alvey Programme, which they couldn’t staff.  And because of their relationship with ICL from the other things that Praxis were doing, astonishingly Martyn recalls, they gave Praxis, a five-year commitment to £2 million worth of work a year.  On the back of it Praxis just started hiring and growing. In July ’83, Praxis had already decided to hire 6new graduates and started teaching them ALGOL 68 and structured programming techniques.

Praxis were very ambitious in what they were doing and kept recruiting into the growth that they could see ahead of them. They just carried on growing, starting off with very few people, and by the end of the first year they had twenty-two or, twenty-four, and by the end of the next year the Praxis team was up to forty-seven staff, and by the end of the year after that they were over seventy.

From his 9 years at Praxis some of Martyn’s proudest moments were: becoming the first independent software house to win the quality standard- British Standard 5750 for all its activities [BS 5750 later became ISO 9001] and Winning the CDIS air traffic control system for, for National Air Traffic Services, which he says was a very proud, and very scary, moment as the value of the contract exceeded Praxis’s annual turnover by some margin “We bet the company on CDIS” and decided it had to be a formal methods project”, he recalls; “NATS later told us that it was the most reliable system they had ever bought”.

Praxis also won the Queen’s Award to industry for their joint work with RSRE Malvern, on ELLA, the VLSI design system.

The success of Praxis, Martyn feels, was due to their determination to be a high-quality software company and focus really on quality ahead of profit. They were very much an organisation that first and foremost believed in treating software development as an engineering process and would not compromise on that. Hence, Praxis had their ISO9000 certification and international leadership in the industrial use of mathematically formal development methods (VDM, Z and SPARK).  This gave them a very strong reputation, that makes Praxis an important footnote in the history of software development.

In 1989 Touche Ross, came knocking on the door looking to acquire a software company of around the size of Praxis who saw this an opportunity to take the message of proper software engineering into a very very much wider space.

Praxis became the computing division of Deloitte & Touche Consulting Group, where Martyn and David Bean became partners in 1992.

Two years later, the Deloitte & Touche partnerships merged their international computing divisions to form Deloitte Consulting. Martyn and David found that the consulting group was dominated by the American partners, and that actually they weren’t in a good place at all, because they didn’t have any control over their own future. He also found that they were part of an organisation now that didn’t believe in the kind of software engineering that Praxis believed in, and indeed, was terrified of the safety-critical work they were doing. In 1994 Martyn was told to close down the critical systems division that was doing all the safety work and where all the formal methods stuff was. Instead he negotiated a sale to Altran, a major French engineering group, and it became Altran UK.

In 1994 after the sale of the Critical systems side of the business, Deloitte & Touche needed somebody to lead the year 2000 services.  Martyn remembers that he hadn’t really realised just how serious a threat Y2K was. But once he had realised, he became a very strong advocate for urgently sorting out the issues, because he could see how, how damaging it was going to be if the problems weren’t solved.  So, he spent some time leading Deloitte Consulting Group’s Y2K activities internationally.

In 1996 Martyn chose to resign from Deloitte & Touche; he recalls that the culture was so different from anything to do with software engineering as he understood it, and the culture within the partnership at that time felt to him to be toxic. It was too money-focused, and there was a level of bullying that he couldn’t tolerate.  Martyn returned to working for himself as a consultant and Expert Witness on large litigations in the UK and Australia.

If Martyn could do things differently he would not have sold Praxis. With hindsight, he wishes they hadn’t done it.  But he says it looked like the right thing to do at the time.  It was a hard lesson to have learnt in a hard way. But Martyn did manage to rescue the critical systems division by, getting Deloitte’s to agree that he could sell it, rather than close it down.

Martyn has in the past been visiting professor at Oxford, Bristol and Aberystwyth universities, and has recently been appointed to a visiting professorship at Manchester University in the Thomas Ashton Institute, which is doing research into reducing risk and improving regulation of the safety space.

Martyn has worked with universities throughout his career, he sees this as a way of getting good technology into practice, which gives universities, who value the commercial experience, the opportunity to work on real commercial problems, and of course Praxis benefited from having access to really leading theoretical computer scientists, to be able to help solve particular difficulties. This relationship Martyn says has been enormously beneficial to him, and it seems right that he should be doing his very best to repay that debt.

Martyn has very broad view of a lot of different technologies in different parts of science, and different parts of engineering, and how they interact, which seems to be very highly valued by academics. He is always pleased and surprised that, that the contributions that he makes are considered, significant by people who he respects for their astonishing intellectual achievements, and academic abilities.

Throughout his career his career Martyn has been on various government committees for different government departments and for the European Commission.

This includes the leading on science for the Board of the Serious Organised Crime Agency where he was able to provide the technical support to the board that was needed with their IT modernisation programme, and other technologies that really needed board oversight because they were so critical to the operation of the entire organisation.

Martyn has spent the last four years on the Boards firstly of the Health and Safety Laboratory, in Buxton, one of the few remaining Public Sector Research Establishments, needed by HSE in order to do the fundamental research in the science that underpins both the policy work in health and safety and also the major accident investigations that go on when something dreadful happens.

Martyn then applied to join the board of the Health and Safety Executive itself. He is still a member of that board and chairs the board subcommittee that oversees all of Science Division’s work, and all the science, the engineering, and the evidence gathering that is done within Science Division for the Health and Safety Executive to provide that fundamental support to all the other work.

Martyn has often appeared as an expert witness throughout the world, he has been involved in litigations with upwards of three-quarters of a million documents to become familiar with in order to be able to look at the real history of a project and to determine exactly why it is that that project ended up where it ended up.

He feels however that to be able to master the complete project well enough to be able to stand up in a court and be cross-examined about it in front of a judge, for days at a time potentially, requires a level of concentration, focus and hard work, that frankly he does not want to do any longer at the age of seventy

From 2015 – 2018, Martyn was the Inaugural Livery Company Professor of Information Technology at Gresham College.

Gresham College has been providing public lectures of high academic standard free to the public since Tudor times, and continues to do so on a regular basis, moving into the twenty-first century by streaming lectures live on YouTube. They have a big digital archive of previous lectures that are freely available at Gresham.ac.uk.

Martyn has delivered 19 lectures in the past 3 years under the title ‘Living in a Cyber-Enabled World’. The thread that ran through all the lectures was that we do live in a digital society, in a world that, that has been empowered and, created to a large extent by the developments that have spun out of the invention of modern computing. He feels we need to take advantage of the enormous opportunities that we’ve got, from autonomy, from robotics, from artificial intelligence, and from software engineering, to make the world a better, more prosperous and a more secure place to live in, if we only do things right.

 Martyn feels, that we have built an astonishing society on foundations that have not been well engineered.  He suggests there is an enormous difference between being able to produce some powerful functionality and being able to produce that powerful functionality to a product quality that makes it fit for purpose in a critical application. And that difference is engineering, which most companies do not regard as fundamental to software development and which universities are increasingly not teaching.

Martyn hopes the interaction with the public audiences in the Museum of London where Martyn has been giving lectures for Gresham College has helped him get the message of the importance of rigorous software engineering across to a much wider group of people than he has managed to reach before.

You can see Martyn’s lectures for Gresham by visiting his Gresham College Page

Martyn thinks that the big challenge is going to be exploiting the enormous opportunities that exist, particularly with artificial intelligence and autonomous systems of various sorts, in a way that is robust against the now really threatening cyber environment, the offensive cyber-attacks.

The work, the strategies that exist, the recommendations that are put forward at the moment, are reactive. They are an attempt to try to minimise the consequences of the defects that exist in the software, rather than looking to the future and saying, how can we replace the software with software that does not contain exploitable defects, and how can we make sure that new software that is being put into critical areas doesn’t contain exploitable defects?  There’s no strategy to do that, yet, and, and no real commercial motivation for commercial companies to do it, because they don’t carry the consequences of, their customers being cyber attacked, or of systems failing and causing societal damage through cyber-attacks.

If we can’t solve those problems, Martyn feels we will run into difficulties where we can’t take the benefits that society demands from the increased use of these technologies.

For Martyn a new graduate faces the dilemma of considering which part of the software industry they want to go into, because, the parts that are really making a difference to society in a positive way are not typically the parts that are going to be making money rapidly.  And when you graduate with a lot of debt, he can really understand that what you want to do is to make money, and so the temptation of going into the City is very great.  He says, “all strength to you, but, once you’ve got the money and you’ve paid off the debts, use your skills for social benefits, and start finding ways to help people to build the systems that society depends on, to a standard that really warrants the trust that is being placed in those systems.  Because that’s hugely important.”

For those who would be interested to look into areas of good practice, Martyn would strongly encourage people to read the seminal papers and books that exist about strong software engineering. But the best thing you can do is to find companies that are building really critical systems to very high standards such as Altran UK, where many of Martyn’s colleagues from Praxis have gone, and continue to build extremely high-quality processes that he has been admiring from afar   since they moved apart.

Interviewed by: Jonathan Sinfield on the 30th May 2018 at the WCIT Hall

Transcribed by: Susan Hutton

Abstracted by: Debs Woodcraft