Tony Storey is a computer scientist who reluctantly accepts the job description “Software Engineer” and has been responsible for some of the most important software developments of the last 50 years, including the ubiquitous CICS, Java and the message-based system, MQ Series.
He started using computers for his work as an experimental chemist, migrated to realtime systems for Ferranti naval weapons and thence to IBM UK’s scientific centre where he helped develop a pioneering relational database system used by the World Health Organisation, among others. He moved on to Hursley and achieved the accolade of IBM Fellow.
Tony explains in this interview with Richard Sharpe how he is driven by the fascination of making things work, often challenging conventional wisdom
Computer scientist, Tony Storey was born in 1941 in Bishop Auckland County Durham, he has one younger brother. His mother was a housewife and his father was a gardener before the war. Tony says he has inherited his father’s genes for gardening.
At the start of the war, Tony’s father joined the Army and was killed in 1944 by a V2 bomb in London. After the war Tony’s mother had to go out to work to keep the family, taking on a number of non-skilled roles from bus conductress to working in a café. His grandmother moved in with the family to help look after the boys while their mother went to work. He says: “We were a poor family, there wasn’t a lot of money coming in.” Tony’s mother eventually remarried.
Tony’s mother was keen on education for her children. Tony says: “She pushed me no end. I went to the local infant and junior school which I loved.”
Having done well in junior school, Tony sat the Eleven Plus and went on to the local grammar school; King James I Grammar School in Bishop Auckland. Initially, he found it difficult to settle and his mother was called in to discuss his future. He adds: “The problem was I couldn’t get interested in a lot of the material that they were teaching us in physics and chemistry. It was really not very interesting and I didn’t get on with it. I was more interested in playing chess in those days, and at age of thirteen I got to the finals of the Durham County Junior Chess Championship. I lost in the final. I’m a very keen chess player.”
Unfortunately, the school did not recognise Tony’s ability from his chess playing, he adds: “They were more interested in the academic side and chess didn’t really come into that. Academically I wasn’t doing particularly well because I wasn’t essentially interested in some of the stuff that I was being taught.” However, Tony managed to find a couple of mentors among the teachers, as he explains: “The history teacher was a very good mentor for me, and he got me very interested in history. I got fairly good passes in eight GCE subjects. Science started being of interest to me big time and I would go to the local library and pick up some advanced books on Scientific subjects; the sort of stuff that they didn’t teach you at the school, and took it on from there.”
After gaining his GCE O Levels, Tony had to leave school to go out to work and earn money for the family.
In 1956, after leaving school, Tony wrote to Bakelite, where his step-father worked, and asked for a job. Bakelite manufactured polymers, PVC, and polyethylene derivatives. His application was successful and he says: “Maybe I showed a level of enthusiasm that most other people didn’t. I had read this book on polyethylene and some derivatives of it. I went along for the interview and blinded them with the fact that I had read the book; none of the people who were doing the interview had, which I thought was highly amusing.”
Working in the Research and Development Laboratory, Tony found the job interesting but wanted to continue his education, so he started a part-time degree, attending the local technical college one day a week and two evening classes per week. He ex- plains; “That was quite hard but fortunately I was doing so well that the head of the science department at Teesside College of Technology wrote to Bakelite and suggested they give me a sandwich course, which they did. That entailed doing six months of the year education full-time in the college, for which Bakelite paid the expenses, and my salary, and then six months at work. I did that for four years.”
Tony studied Physical Chemistry, organic chemistry, inorganic chemistry, and maths. He graduated as a Member of the Royal Institute of Chemistry with Honours. He was also awarded the ICI Science Medal, which was awarded annually to the most outstanding science and technology student attending the college at that time.
After graduating, Tony was invited to complete a PhD. He says: “After my degree finished, my organic chemistry teacher, who was a huge influence on me and the next stages of my career, said, ‘How would you like to do a PhD?’ I hadn’t really much idea what that involved but it sounded interesting.” Tony’s teacher introduced him to Professor Musgrave, professor of the chemistry department at Durham University, who invited Tony for an interview.
Tony says: “We met in a pub in Durham and chatted a little bit. I was offered a grant from the Imperial Smelting Corporation to do three years doctoral research in organo-fluorine, heterocyclic fluoro organic compounds, which I knew absolutely nothing about at the time, didn’t even know they existed. In fact they didn’t exist, because fluorine is not a naturally-occurring element in hydrocarbon compounds so, it was all synthesising completely brand new chemical structures. I was absolutely fascinated by this whole thing, it was just such a wonderful experience. After eighteen months I had enough results to write my PhD thesis which was 250 pages long.” Tony’s course was for three years, so he was given additional projects to work on.
It was while working on his project that Tony began using computers. He says: “It was during that time actually that computing crept into the subject. When you synthesise these organo-fluorine compounds, and perform chemical reactions against them, you have no idea what the end result will be because it’s brand new stuff; nobody’s ever done this before. … The primary weapon I used was nuclear magnetic resonance spectroscopy because fluorine and hydrogen are two elements that react very strongly in electromagnetic fields and they absorb radiation producing spectra from which you can deduce roughly what the possible structures of the resulting new compounds were. But to do that, there were some ambiguities which had to be resolved.
“So I chatted to another one of the Readers and he suggested that I try and work out the electron densities of all of the carbon atoms in these molecules which could tell me where the fluorine atoms all were. To do that I had to solve some complex maths problems and use a computer to do that After a struggle, I wrote a program. They had an Elliott 803 computer at the university and a KDF9 at the University of Newcastle which we could use. The programs ran successfully and I was able to pretty much work out what the structures were purely from the NMR spectra that I received. As a back up, I had to break those new molecules down in certain ways, such as oxidation, to something simpler that was a known fluoro compound, and that verified, eventually, that my calculations and my assumptions about what had happened were correct.”
Tony wrote his program using ALGOL 60. As a result of the involvement with computers during the last year of his PhD, Tony become fascinated by computing as a subject and was torn between his love of chemistry and his new found interest in computing.
At the end of his PhD, an opportunity opened up to do some post doctoral research at the University of Southampton using liquid crystals as the medium in which NMR spectra were taken. This allowed more accurate determination of molecular structures from the resulting spectra. He applied to the Department of Science and Industrial Research, and received a grant for two years work, during which time his interest in computing continued to grow. He says: “The interest in computing grew ever stronger, because, again, to analyse results you still had to use computing programs to solve some of these problems. I wrote programs in ALGOL I used to go to the famous Atlas 1 machine at Harwell and use that for some of the more complex problems. They had a local ICL 1904 machine at the university and I got well in with the computing department there. The work on that was mainly batch-oriented; you put stuff in and waited for the results to come back. It was pretty clunky, and they were trying to write a new piece of software operating system which made it much more amenable to use. I got involved in some of that indirectly.”
At the end of his two year project, Tony decided he had to get a job and applied to Reckitt and Colman who wanted a person to do NMR for them. However, at the interview, Tony found that the job sounded too narrow for him. At the same time, he saw numerous adverts for people to work in computers and decided to “have a go”.
He applied for a number of opportunities and was offered interviews for all of them. He attended all but found most of them failed to interest him, he adds: “I walked out of the one with IBM. I left halfway through the interview because it sounded pretty, pretty awful to me. The one that attracted me most of all was the one at Ferranti, Digital Systems Division, in Bracknell.”
Ferranti, Digital Systems Division
In 1969, Tony joined Ferranti’s Digital Systems Division in Bracknell which focused on real-time systems. Tony says: “This was something completely new to me. It sounded interesting and it turned out it was building weapons systems platforms for the Type 21 frigates as they were at the time, and the Type 42 destroyers. The computers were fascinating, they were absolutely amazing. I had never seen anything like them. They used a three-address instruction code. You could do three operations at the same time in one cycle. So that meant they were very powerful.”
Tony says: “When I started there were a whole group of new graduates starting and we all went through a training process together and all lived together. It was just like being at university but a bit more intense. The training at Ferranti set me up for the rest of my career.” The training involved working with an autocode and manipulating registers which gave Tony a deeper understanding of how the machines worked.
I also read many books and papers software engineering. It was a hugely immersive place to work.
He adds: “I was tasked with working on the basic operating system for the machines. It was a 24-bit word machine, so it was quite an unusual word size. It was a 16K ma- chine, so you had to write the operating system in about 2k words. This was dealing with lots of real-time devices, memory and storage management, multiprogramming, fault handling and more, so it was a difficult task to keep it down to that number of words.
“At the same time I was also given the job of creating a time-sharing system for them on this machine so that the various people who were doing the development could have terminals directly connected at the computer and do their own thing on that computer. That was a very difficult task.”
During his time at Ferranti, Tony became team leader of a small group working on the multitasking time-sharing systems. He says of his management style: “I fostered teamwork. Everybody was equal. There was no hierarchy. It was just encouragement and being enthusiastic. If people had problems, they could come and discuss them with me. I couldn’t guarantee to solve the problems but at least I was there to listen to what they had to say. I listened to their input quite a lot. I have to say I hated the idea of being a manager. It’s not really my scene, I’m afraid.”
In 1976, when the Division was moved to Wales, Tony decided not to follow and after seven years, he decided to move on. He says: “I had no idea what I was going to do but I saw an advertisement in the Guardian one day, and it was for somebody to work at IBM, at the Scientific Centre in Peterlee, County Durham.” Tony applied, had an- other pub based interview and was offered the job on the spot. The role was in the emerging field of relational database technology.
Tony explains: “The theory had been proposed by Ted Codd, an Englishman, who es- caped to California. He proposed this set-oriented approach for managing data. The IBM Scientific Centre was a small place in Peterlee and they were tasked with the idea of trying to develop a language to express the various relational operations, and a system that could run, store, retrieve the data; did things like optimisation. I was particularly interested in the field of trying to synchronise multiple users across databases at the time so I worked on that. It was another kind of academically-oriented thing.”
He describes the culture as free and easy and more like a family than a workplace. Tony’s work saw him develop the system to allow queries against large quantities of data to happen efficiently. He says: “It was also trying to extend the language to allow for certain new algebraic operations. We built a prototype; the Peterlee Relational Test Vehicle, PRTV. We tested it in a number of places, and the big test was a joint project with the World Health Organisation (WHO) which had gathered large quantities of statistics and information, from all over the world, about mortality and morbid- ity statistics in various sexes and age groups and so on and so forth. … We also had a lot of interest in it from agricultural institutes from around the world, people in South America studying breeding of potatoes and others studying wheat.”
As a result of the early success, Tony spoke at conferences, gave lectures and visited customers. He adds: “It was a good time.”
Asked about the culture of IBM, Tony explains that the Scientific Centre, was some- what separate from the mainstream so a lot of the IBM culture did not ‘show through’. He adds: “It was well out of the way up in the north of England. Although there were IBM offices in Newcastle, Manchester, Leeds , they were sales-oriented for the main- stream products that IBM was making like the 360, and the various software that went with that. I would hesitate to call it a skunkworks, but the scientific centre was bordering on that. Everybody believed at the time that what we were working on was the future for database technology and the guys in IBM Research in Almaden were work- ing on System R at the time which was very similar kind of thing.”
On the durability of relational databases, Tony adds: “I think it’s something that’s extensible, various people have produced various extensions to it over time and it morphs into something slightly different, with a slightly different focus, but it’s still a set-oriented relational technology under the covers.”
In the competition with Oracle who were also working on relational database techno- logy at the same time, Tony says: “I have to say, that IBM was pretty slow on picking up on this relational technology, largely because they had a database product called IMS which was like a hierarchical database technology. They had lots of customers using that. So they were very slow in pushing relational technology into the market. In fact they were behind Oracle into the market who got the first commercial implementations of relational database technology out to customers. IBM were then running to catch up and they eventually came out with DB2. It was a struggle all the time. Every time you mentioned the word relational, you had the IMS guys jumping all over things. It was a sort of culture war.”
In 1981, Tony moved to the IBM research and development laboratory at Hursley where he started working on the design, and development of an online transactional processing system; the Customer Information and Control System (CICS). CICS al- lows secure and proper transactions to occur and if a transaction fails, it rolls that par- ticular transaction back to where it was before it began so that it can be repeated. Tony explains: “It dealt with the whole issue of data integrity, ensuring that the data was reliably maintained in a consistent state. It was basically a large amount of soft- ware which sat on top of the basic operating system The operating system did very little to help where you wanted to perform short request response interactions to manipulate a particular piece of data, or collection of data, and do it quickly and reliably and with high integrity. Throughtput and scalability were key. So we built this thing. It was initially built as an application for a utility company in the US. It was transferred to Hursley developed over a period of many years. In fact it still is being further enhanced to incorporate current developments in software.
The software has made billions of dollars in revenue for IBM, Tony adds: “It has be- come the de facto standard for online transaction processing. … It’s used by enter- prises all over the world to run their businesses and so it has to be reliable, it has to always be there. It has not got to fail, and if does fail, it’s got to be able to recover quickly, and with a high level of integrity bestowed on it.”
As well as running it on IBMS’s proprietary systems like the 360, Tony was involved in the transfer to open systems and IBM’s AIX. He explains: “I, with a lot of initial resistance, decided I wanted to put this product on the RS/6000, which is a RISC computer that IBM came out with and it ran Unix, or AIX as the IBM version of UNIX was called. Again, that was a bit more of a skunkworks. I got permission by my manager at the time to give it a go. We spent some effort and time and got it running on the RS/6000 under AIX. It was a complete reimplementation of CICS, exploiting some of the specific hardware features, like the memory segmentation of the RISC/ 6000 and we were able to take that and transfer it to other platforms.” The software also eventually ran on HP, SUN and DEC machines and was eventually called TX Series.
Tony went on to propose and develop the IBM family of messaging products known today as WebSphere MQ. He explains: “It’s two things. It’s a way of delivering mes- sages reliably between programs. But the important thing is that the recipient doesn’t have to be around at the time. So it can be delivered later on but still reliably. The message is managed all the way through the cycle of being sent all the way through to somebody at the other end who receives it. It also presents a new programming model for building applications, where you can take an application that exists for ex- ample, and build a new one and connect…You don’t have to modify this, but you can connect this other thing to it through this message-driven processing technology. It’s reliable, guaranteed. It sounds simple, but there wasn’t much software around that did that. Simple ideas are usually the best ones.”
Tony says that initially it didn’t gain as much traction as CICS, but that it has become widely used and a foundation for message brokers. Ir was awarded the 2004 Mac- Robert award by the Royal Academy of Engineering, the first software product to do so.
Tony was also part of a collaboration with other IT manufacturers, who worked on developing an enterprise Java infrastructure. He says: “We defined a component mod- el for Java components and various specifications that cover the types of transaction processes, the online transactions, the messaging technology, the data technology and so on and so forth, which were necessary to build enterprise applications with, at that level. So we basically extended the language to build this higher level infrastructure that allowed construction of business applications.”
In doing this, IBM shifted from proprietary approach towards working on international standards that allow systems to work with each other more effectively than they had done in the past.
Tony remained at IBM for the remainder of his career. He was made an IBM Fellow in 2002 and worked for five years beyond his retirement date, finally finishing in 2009. He was elected Fellow of the Royal Academy of Engineering in 2002, and awarded the BCS Lovelace medal in 2009.
On the topic of recent cyber breaches such as Microsoft etc, Tony says: “I don’t think we pay enough attention to this whole business of cybersecurity. It needs a lot more effort on it than is currently being put on it. The problems are giving IT a really bad name.”
Tony says his biggest mistake was being too dogmatic. He adds: “I suppose it comes from a desire to succeed, and not trusting other people, which is a bad thing. I think, I was probably too forceful in my approach to making things happen. I think often you have to be but, you can probably much more diplomatic about it than I often was.”
For those considering looking at a career in IT today, Tony says: “If I was going to do it again, I think I would go into the cybersecurity space. There’s a huge challenge and huge opportunities to make your way in that.”
Another area that interests Tony is quantum computing. He adds: “I think I would probably be interested in trying to get into that field if I was starting all over again. It could be one of those areas that’s always coming, but it’s one of those fields where you have the opportunity to make it work, and make it big.”
Interviewed by: Richard Sharpe
Transcribed by: Susan Hutton
Abstracted by: Lynda Feeley