Problem-solving pushes boundaries
It is well known that equations and algorithms used in rocket science are used in Wall Street, but that is not the only crossover between hard-core scientific computing and the commercial world. Scientific research progresses because researchers constantly come up with new ways to meet the challenges thrown up by their research goals.
The Numerical Algorithms Group was founded on the idea that what can solve a research problem for one scientist might solve another problem in a completely unrelated field. It has amassed a huge library of algorithmic software that sometimes is obtained from research or sometimes it develops itself. It has no shareholders, just members. If it runs out of money – that's it folks. While to a certain extent it has a public service mission, its prime mission is to keep paying the bills so it can continue to produce software that solves the next generation of problems.
After more than 30 years of successfully updating and selling this library, its founder retired from the front line. Into this very British academic world came an ebullient American, Rob Meyer. Although he had a foundation in scientific computing, his background was much more business oriented.
But he recognises that NAG has great strengths and wants to build on that, while also using his marketing skills to find new opportunities that will keep this rather unusual software business buoyant well into the future.
David Sayers, principal technical consultant at NAG, says that Meyer not only has business skills, but has enough technical knowledge to gain the respect of both technical customers and the development team.
He says: 'He is very easy-going. His management style is to liberate them to be creative. He pulls it all together and co-ordinates things, but he has allowed individual flair to flourish. If we have good ideas, then he tells us to give it a go.
'What impresses me most about him as a director is that he knows how to use our products. Also, he has impressed me because he knows what he doesn't know. He will talk knowledgeably about something and then say “this exhausts what I know and I don't think I can contribute to the discussion any more”. I think that is a great that he knows his limits rather than just carrying on.
'He has made an immense contribution, because we are all now more relaxed and willing to work for the company. He has given us a great sense of unity and made us feel that we are all pulling in the right direction.
'Before he came we tried to do too many things, and if you over commit you end up failing on a broad front. We all feel secure and that, whatever we do, it's not going to be a huge gamble.
'His particular interest has been in the marketing side of things. His main contribution there has been unifying the world of the three hubs of the company, in the UK, US and Japan. We used to work a little bit independently and he has got us to think as a single entity. The marketing is starting to come together so we are not duplicating each other's work.
'Nobody minds that he is an American. We are gradually training him in the art of quiet English understatement. I was a bit surprised that he did not move over here when he was appointed, but he does spend a lot of time in the UK and with the wonders of electronic mail, it does not seem to matter. He travels a lot anyway and is a great ambassador for the company.'
Meyer was born in Texas City, an oil refinery town on the Gulf Coast of Texas. His ancestry was German farming folk. One side of family emigrated during the 1840s having previously emigrated from England to escape religious persecution during Tudor times. He has an ambition to trace the full history of his family.
His father worked as a technician at the oil refinery, and was a Marine Corps veteran of the South Pacific war. Meyer and his father were both great outdoorsmen and spent a lot of time hunting and fishing, mainly around a farm that was in the family. Meyer says his father was extremely clever and would have gained a scholarship to study mathematics if he had not been called up after Pearl Harbour.
Meyer was a good all-round student, playing sports and studying – and doing well at both. His best subject was maths, but he was also very interested in science and technology. Computers were just becoming commonly used during his later years at high school and he was fascinated by them. He wanted to study computer science, but at the time such courses did not exist. So, in 1970, he went to Washington University in St Louis to study applied mathematics – mainly because it was one of the few courses available that included computing.
He had won a competitive full scholarship, but it could only be used if he went to college in Texas, so he had to give it up to study in Missouri, but he wanted to do that course. He did get a partial scholarship, but also had to do administrative jobs in the university to help pay his way. He often reflected on his decision to pass up his scholarship as he struggled for years to repay students loans.
He soon developed an interest in transport planning and, after he graduated, moved over to the school of transportation and urban planning to do his further degrees.
He says: 'I had always been interested in urban planning. Transportation systems engineering sounds further afield from applied mathematics than it really is. It's a discipline in civil engineering, but it involves mathematical modelling of transport systems for public works. My research involved reducing those mathematical models to computing solutions. When I came to NAG seven years ago, I was really returning to where I had come from in my academic career.'
He started looking for a job before he left graduate school in 1978, because he was fed up with being a labourer working as a research assistant for principal investigators. His first job was with a company that ran a fleet of boats that towed barges around the river systems of the US. They wanted him to design a computer system to help them plan and operate. It was not the best paying job he was offered, but it was the one that interested him most and he has spent some time working on tow boats when he was a student. After developing the system, he moved up through the management structure and was eventually put in charge of running part of the company.
He had developed a taste for management and, after eight years, the company was sold and he didn't want to work for the new owners. He went back to St Louis to work for a small company that did manufacturing for the aerospace and silicon wafer industry. He was only there a year, but he learned a lot about business. He was then recruited to go back into marine transportation and he joined Ingram in Nashville, Tennessee.
He says: 'That was probably the worst career move of my life; I was very unhappy there. My boss was very difficult to work for. He hired somebody as president over my head and I guess it was a “bait and switch”. Eventually we agreed to part ways and I decided I wanted to do something completely different.'
He moved to a job with the Gas Research Institute in Chicago, which was responsible for planning and managing a $200m R&D program for the US natural gas industry.
He says: 'That was a wonderful opportunity for learning; every day I was drinking from a fire hose in terms of new experiences. I got to learn about intellectual property law and licensing and learned how R&D works and how to manage it. In 10 years, I had seven different jobs. They got me at a good price and kept offering me new responsibility and eventually I was general manager of one of the business units.'
NAG opportunity knocks
He stayed for 10 years, gaining the equivalent of an MBA from the Kellogg Institute along the way. He started to think about again doing something new and, as he was starting to get older, realised that he would probably be choosing what he was going to do for the rest of his career. At this time the opportunity arose to work for NAG in North America.
NAG was founded in 1970 as a collaborative venture between several leading UK universities, led by Dr Brian Ford, a renowned pioneer in the history of computing. Ford's Mentor was Jim Wilkinson, who worked on the Pilot Ace computer and the National Physical Laboratory in Teddington, and was Alan Turing's assistant. It was initially known as the Nottingham Algorithms Group, after the university where it was based, and was funded by the UK government's Computer Board to produce a software library for common use. In 1976 it was made to finance itself and relocated to Oxford. Today it is still structured as a 'not for profit' organisation, but apart from that operates like any other software company. It sells licenses to its vast library of algorithms and uses the profits to fund the development of new ideas and does not pay dividends. Nobody really 'owns' it, but if it were ever to be floated on the Stock Exchange, the UK government would probably stake a claim.
NAG had a presence in Chicago due to long historical links with the high performance computing community around there, particularly at the University of Michigan and Argonne National Laboratory, where Jim Wilkinson spent a lot of time. Within a short period Meyer became head of the US subsidiary, NAG Inc.
Meyer believes they chose him to replace Brian Ford because firstly they wanted someone who would commit to the organisation for the long term, and also wanted someone to 'expand its vision' and transform NAG into a global organisation.
He says: 'In selecting me, they chose someone who was familiar with what NAG did, but came from a slightly different world than the one NAG grew up in and could help bring NAG forward and broaden its horizons. The ideas I put forward appealed to them, despite the fact that I spoke the wrong kind of English and lived in the wrong place.'
Meyer does not believe he has radically changed the way that NAG has operated. He appreciated its technical strength and the strength in depth of its library. His focus has been on the customer-facing end of the organisation, making its products more accessible and adapting to the changing way that people use its software. In particular he has pushed the idea of partnership with other software companies and getting them to incorporate NAG software into packages rather than selling directly to end-users.
He says: 'There was so much potential there. So many great people, great technology – this was real craftsmanship. It had been a great market-leading company throughout its history, but it needed bringing into the modern era. It had taken a turn that was a distraction really, because they thought the library would eventually go away. We thought the world would evolve into a set of desktop tools for solving problems.
'There were a couple of ventures where we didn't really understand how people would be using our software. These became a distraction from the work of continuously developing our library and continuing to understand how commercial organisations would eventually use the software we produced.
'When I arrived, they had realised this, but it was not clear what they were going to do. There was already a partnership arrangement with a company producing mathematical software, so I looked for ways to reproduce this with a number of other software companies in financial and consumer markets. I dramatically expanded our programme of getting our products into those produced by Oracle or PeopleSoft. Previously we had assumed that all our users would be Fortran programmers. That world was disappearing. In the science world, products like Matlab were coming along, so people did not have to be Fortran or C programmers.
'What we realised was that we had unique expertise and something truly good, and if we could communicate that properly and continue to advance it, we could reach a much larger group of users out there. We can either go through large companies or software companies that use our code to add capabilities for people who are not Fortran programmers.'
NAG's original mission was to develop algorithms for use in research, but as the world has moved on, it has found much more fertile furrows to plough in the shape of financial services. Hedge funds in particular are known for using 'rocket science' to create derivative products and plan strategies. Data mining is also an essential part of customer relationship management systems even though the algorithms were invented for more esoteric applications. Many of the libraries of algorithms were torture-tested in other demanding scientific applications before being incorporated into new markets as diverse as retailing and HR.
As most of the partners for this work were based in the US, Meyer found that he was leading the drive into these new markets. Having begun with NAG as a vice president he was soon put in charge of the entire US division of the company and, when Brian Ford decided to take a step back, he was chosen to be his successor as CEO of the whole company.
Meyer is very comfortable with the 'not for profit' status, because he believes that it makes it possible to develop software for the long term, rather than worrying about reporting quarterly results to Wall Street. He does not believe that this means it should only develop software for obscure scientific research, because history has shown that something developed for one application can often find a use in an unrelated field.
He says: 'The only way you advance the frontier for software is to serve all comers. We are in the business of advancing the frontiers for a lot of people's benefit. We are going to take some of what we earn and use it in other projects, such as LAPACK, to give back to the community. In terms of pricing, we charge according to the commercial value of software to an organisation. When it comes to academic customers we charge a lot less because we want to make sure there is a new generation of knowledgeable users who are getting good answers by using this kinds of software. In some cases we give software free and we contribute to open-source projects. We charge government labs less than commercial customers and academic customers a lot less. Our culture is to look at the long term. It is good for us to have people going out into industry able to use NAG software.'