Interviewee: Peter Moforth Job:? Organisation: The Turing Institute Interviewer: Thomas Haigh Date: Wednesday 19th April Location: The Turing Institute Revision: 0 (No feedback obtained from interviewees)
The Turing Institute is named after the great English mathematician and pioneer of computing theory Alan Turing. Turing is well known for his war time work in Bletchley Park, and for his time at the universities of Cambridge, Princeton and Manchester. The link with Scotland comes via Donald Mickey, a colleague of Turing's who became interested in Finite State Automata and was instrumental in the establishment of the pioneering department of Artificial Intelligence at the University of Edinburgh in 1963. Following the Lighthill report in 1973 and the effective removal of government support for AI in Britain, work in the area subsided until the early 1980s and the revival of interest in the subject brought about by the Japanese 5th Generation project. The resultant infusion of money to the field, governmental and industrial, led Mickey to move from his small Machine Intelligence Unit at the University of Edinburgh to a new body called the Turing Institute which was intended to 'sit part way between industry and academe', formed in 1983 as a limited company. Glasgow's status as a development zone made it an attractive location for the new Institute, and close links were formed with the nearby University of Strathclyde. The Institute grew to employ around fifty people by the late 80s.
Scottish Enterprise, who were responsible for the influx of electronics manufacturing to Scotland and were keen to encourage the software industry, stepped in to rescue the company after it ran into acute financial difficulties. The influx of cash was accompanied by 'professional managers, who knew a lot about business and marketing, but they didn't know anything at all about computer software and our business.' Two years, and a great deal of money later, the company floundered in this commercial incarnation.
It was saved by the intervention of Glasgow University, and returned to its core business and original non-profit status - funded by the Glasgow Development Agency and the University. It is intended to cover its costs without further subsidy, and derives roughly half of its income from contract industrial R&D, the remainder coming from 'soft' sources such as ESPRIT and LINK. Since its reformation in July 1993 targets have so far been met and an upturn in business has taken place. The University has made it clear that more money will not be forthcoming and that their loan will eventually have to be repaid.
'At the moment about 60% of our turnover is commercial, where we deal with companies, and the reason why we are in a position to win business quite successfully is that one of the things that the University does is to very significantly lower our costs as a commercial company.'
These advantages include cheap accommodation, low cost network access and IT infrastructure and access to graduate students. 'We're a small company - at the moment we only have seven members of staff, although there's quite a number of PhD students.' Their links with the university give them advantages comparable to those enjoyed by members of a huge corporate research lab, such as access to experts in a vast number of fields, but with none of the huge cost overheads which such organisations entail for their parent companies.
'Depending on where you're coming from you can see the organisation in different ways. One way of reading the articles and memoranda is that we are a part of the University, and that is very important because it means that success that we have in winning R&D moneys, in publishing papers and running PhD students are counted within the University's success, and as such they get a better score in comparative studies with other universities... and they get a larger block grant. That's the main benefit that they see out of it.
As far as we're concerned we have this access to cheap labour and cheap infrastructure, it's symbiotic - quite mutual. However, when we talk to commercial companies it is very important to stress that we are an independent, commercial company and we tend not to stress the university involvement. I'm not talking confidentially here, because a lot of other organisations do this, but when we talk to suppliers about getting equipment and they say "Are you part of the University?" we say "Oh yes, we're part of the University", because you get your academic discount. There are other sums of public money which you can only get access to if you're a commercial company - universities don't have access to that, we do. Basically you're dealing with an organisation which has a big hat rack: some commercial and some academic. Every new situation that comes along you adjust your presentation to optimise the amount of resources that you get. It's very simple.'
The work of the Institute has machine learning and knowledge based systems as one of its main foci. Their approach is firmly symbolic, involving the automatic generation of humanly transparent production rules from data. For business use this technology has the distinct advantage over sub-symbolic approaches like neural nets that the operation of the system can easily be understood, debugged and validated - important for mission critical applications and for compliance with various international standards. Their other main area of activity is 3D image processing - in the last year they have serviced four separate contracts involving the synthesis of a 3D surface image from photographs. This technology has a wide range of applications - from the production of 3D models of the earth to the representation of human heads for medical work, and such has been the interest in this area that they are currently working to 'productise' the system.
This is a shift away from their mission as a contractor of R&D.
'It's an interesting question - how often do you do repeat business before contract R&D become a product. If you do the same bit of contract R&D twenty times then it begins to look very much like a product. One way to migrate a technology from contract R&D to being a commercial product, which is a completely different business with almost nothing in common, is to do the same thing over and over again. I say that because the contract R&D business has a culture and a community of people in it whose interests are in "pushing the envelope", they have academic and scientific interests to "boldly go" - whereas in a products company everything's about the bottom line, and it's basically 99% marketing and sales and 1% yesterday's technology. It's likely that our organisation may split in the coming years, because of the cultural problems of keeping the products side in parallel with contract R&D.'
In his view it is inevitable that universities will eventually 'have to bite the bullet' and consider the setting up of spin off organisations to provide support and maintenance on any products which they wish to commercialise.
He sees the move by larger companies towards 'rightsizing' and increased reliance on external contractors as working to the advantage of service organisations, including research service organisations such as the Turing Institute. 'It's definitely flowing with the direction of what's happening in the marketplace - people outsource and you go in there and get business which otherwise would have been handled within the company.... People have come to recognise that the most successful companies are small, but make up for having a small size by having a big network.... The social side is as important as actually writing code'.
Does he see technology exploitation in the software field as similar to established methods in other fields, such as biotechnology or pharmaceuticals? He holds the film industry up as an perhaps the closest relative - like software it is project based and has technical and non-technical components. Studios are loosely organised, large numbers of small and highly specialised companies handle work on a contract basis and many participants are self employed. Distribution is a separate process and can be handled via a variety of channels. 'We're probably where Hollywood was in the 1920s - still at the silent movie stage. Perhaps in the 30s.'
Have traditional ideas of basic research, applied research and product development no place in the software field? The film industry model suggests that the strict framework of conventional science, funding bodies, laboratories, hypothesis and so on with research papers as the primary output is restrictive. 'You don't do software for its own sake, software is always targeted to somebody else's problem. That always takes you into other disciplines'. Within the Turing Institute itself some have artistic backgrounds, others were trained in mathematics, physics or computer programming.
'The traditional departments which make up universities, the Department of Physics, the Department of Chemistry, these things are starting to break down anyway - and what you're replacing it with is just a big soup of people.... You work in one subject and you end up contributing to another subject - and all the time, when dealing with people in the university doing mechanical engineering and genetics and geology, all the projects that people are getting involved with are all interdisciplinary, multi-disciplinary.... I think that because computers and computer software are so esoteric that they tend to permeate almost everything so that to an extent people in the area of computer software tend to be leading this cultural change.... replacing the conventional barriers within universities with a heterogeneous series of barriers, all of which are project based.'
For example, a robotics initiative within the university also involves the departments of zoology, aerospace engineering and mechanical engineering. These kinds of temporary groups of interested participants, raising money and working together for a particular purpose, again remind him of the film industry. This kind of organisation is far more suitable for the addressing of complex, real-world problems - 'I can't think of any problem we've come across in the last five years where you can go to a single department and have all the expertise you need.'
How important does he feel local contacts are? Despite the potential of modern communication technologies, 'none of these are a substitute for people getting together, scribbling on the back of beer mats and that kind of thing.' Similarly, although bands can now record together without entering the studio at the same time, the creative quality of the music suffers. 'I think geography is very important - location and image'. The Turing Institute is just a five minute walk from the campus - distinct enough preserve a separate identity, but close enough for regular social interaction and university members to drop by for coffee. In future the role of a computer lab may become primarily social, with development work best undertaken in more private conditions on a laptop or at home. However, so far no particular benefits have accrued from the relative concentration of computer companies in 'Silicon Glen' - 'my personal view is that we are yet to achieve the critical mass' achieved in California where high staff mobility and strong personal contacts ensure rapid transfer of ideas and expertise.
'All the time... I keep going back to social things, cultural things and presentational things. I think that those are some of the most important issues to get right if you're going to treat software as a business successfully and make money from it.'
The Institute has been involved in projects as part of ESPRIT I. II and III. How did he regard the experience?
'I regard it as a total a waste of time. We were in there right at the start, and we've been involved in a dozen different ESPRIT projects in one way or another, and I can't think of a single useful bit of technology we generated, or a case where we did something that led to any kind of genuine commercial business or a technology produced as an ESPRIT result actually went out and was commercially successful for Europe. The only beneficiaries I have actually witnessed are the international airlines and the hotels. I'm extremely cynical....'
The projects involved work with organisations of all kinds, including large and small companies and universities. 'I suppose we have made some interesting links, but I honestly there's any business has come from them.' However, a further ESPRIT proposal is currently under preparation. Preferred projects are those 'in the corners' of a diagram which are not reliant on the inputs of other groups.
'It really has to be a piece of code that you've already written, or failing that - which would be annoying - something that you had to do anyway.... You've also got to make sure that there's a member in Greece or Portugal to hold your summer management meetings and someone near Grenoble for your winter meetings - quality of life. Given the amount of money that's gone into ESPRIT it would have been far better used if people had made these projects much smaller, with a much much lower bureaucratic overhead.'
What is the Turing Institute really in existence to accomplish?
First and foremost, operating as a hard-nosed commercial company, to make money. If you've got money you can do anything - make a nice pleasant environment for yourself to work with, buy lots of machines, pick and choose on some things, it's definitely a way of solving a lot of problems.... The second element of the mission statement is "to boldly go" - everybody inside the Turing Institute has got a reasonable CV. I've got about 80 research papers to my name, and although I've talked about the commercial side you always like to carry little grains onto the beach of scientific knowledge... doing a bit of that is pleasant. The University can build that into their metrics of success, so there's no conflict there. The third element, which is implicit in almost every organisation, is doing things so that you can hopefully have a bit of fun... the Turing Institute organised the first Robot Olympics in 1990 - that was a real boost, put us on the map and it was a lot of fun.
How does he react to the suggestion that universities are public institutions which should not be heavily involved in ownership or support of commercial companies competing in the marketplace?
I think that is nonsense. If the suggestion here is that private companies are not receiving soft money from the government then that's baloney - just look at Nissan, how many millions did they get? You look at all the grants and so on which are available to commercial companies - we've set ourselves up sitting on this fence to try and kind of milk from both sides, and if you've set it up right and have a clear view of your goals and objectives and are prepared to do accounts and all that then you get your VAT back - that's 17.5% on the bottom line we're talking about, and there's all kinds of things you can do as a company which connect you to soft money which universities don't have access to.... Universities should be free to compete, the only thing is they need to be a little bit more professional about it, and a lot of that just comes down to people being honest with themselves, I think, about what they have to do.
I think universities do need to undergo some restructuring in thought, in as much as part of what a university is there for is to be a centre of broad based expertise in research and development... a conglomeration of people with in depth, specialist knowledge in certain areas. A university should also be a place where large numbers of people come to get taught, and I think that those are the two vision statements of universities, and that duality should be reflected far more in the physical structure of universities.... This obvious and natural dichotomy of what the goals of a university are will separate out into the R&D thing, in which people need to be sufficiently professional that they can compete on equal terms with commercial companies, and the teaching side - with whatever metrics and so on are appropriate there....
Whilst the two tasks should remain in the same institution, there is a case for differentiating between teaching and research staff, especially since prestige and promotion tends to accrue only as a result of research and publications under the current system - which is unfair to staff whose interests tend towards teaching. 'I think that's something which universities need to come to terms with, and they have barely started that process.'
'If you look at many computing science department they run like businesses - take away the teaching side, what you've got left is at the level of a commercial business. They've got millions of pounds worth of turnover, they've got contracts, they've got deliverables, I think one of the major reasons why universities do find it hard to compete is that they are managed by people who have so little experience of those things. Typically you find the senior manager is an academic with a great CV who is suddenly promoted to be the departmental Prof. and are expected to play the role of a company CEO, and I don't think that they're cut out for that. There are some departments which are headed by people who do have the skills of a company Chief Executive, they are few and far between, but boy do they make their departments go super nova... they can get incredible leverage over the departments which aren't as good as they are.... Sometimes departments bring in people called Business Managers, but when you look at how this is done, particularly in British Universities, they do what the Prof tells them.... On the business side of things, and let's be blunt it typically is a business - look at all the contracts you get - they should be run by someone with formal business training. Most Profs don't even know how to read accounts... and I think universities lose out by not having a management structure that reflects what they're trying to do.'
He doesn't see any conflict between a more 'commercial' approach oriented to fund raising and the continuence of a strong tradition of basic research. Rather increased activity will create a surplus which can be chanelled to knowledge creation.
'Look at MIT, it's very comercially managed, yet you've got lots of people in there doing totally hare-brained things. The important thing is to get away from all these horrible old monastic traditions which I think have held academia back.'
'All of these ideas are only relevent at this time now - I'm sure that if we'd been having this conversation in the late 1980s I would have said different things, and all the people you've talked to would have said different things, and in five years time they'll be saying different things again. I've used the example of the football team - because these are the rules which apply now we're all developing these finely tuned strategies to optimise our play. At the moment it's football, maybe by the turn of the century it'll be hockey and we'll all be playing a different game.'