Universities are also being exhorted to inculcate business skills into their students. The recently published "Dearing Report" [3] into Higher Education in the UK not only contains recommendations about including work experience in degree courses, but also (Recommendation 40) "We recommend to higher education institutions that they consider the scope for encouraging entrepeneurship through innovative approaches to programme design and through specialist postgraduate programmes." This recommendation arises in part from an appreciation that the traditional route into industry, through employment in large companies with graduate training programmes, is no longer an option for many graduates. Many join small or medium enterprises where they are expected to contribute immediately to the company's commercial activities and there is also a hope, at least, in government and commercial circles, that some graduates, and particularly higher degree graduates, will set up their own companies following graduation.
These kinds of expectations are particularly high in the area of information technology, which has been a strong growth area for many years, and which is expected to continue to grow in potential and importance for many years to come. Furthermore, there are examples of spectacular success by start-up companies in this area, particularly in the USA. Direct exploitation by universities of the results of research in IT is not straightforward however. Intellectual property rights also carry responsibilities and this is a particularly significant issue in IT research, where the outcome is often manifest as a piece of software. Questions of maintenance, product liability, etc. then become major concerns both for the university and the individual member(s) of staff involved as the `inventor(s)' of the software, as do the divisions of rewards and responsibilities between the university and its employee(s). Both parties may consider that it is better to give the software away, as `freeware', thus avoiding the potential liabilities and hopefully achieving fame, rather than risk these liabilities by selling the software and thus, potentially, making a fortune.
Both undergraduate and postgraduate students in research-led universities who join industrial companies frequently take with them knowledge and expertise gained whilst undertaking projects as part of their educational experience. In the UK there are also formal mechanisms to support this process such as Teaching Company Schemes. The Teaching Company Scheme partially funds recent high quality graduates to work on company agreed projects for up to 2 years, during which time the students have both a company and an academic supervisor. This mechanism has been successful in both transferring technology and enhancing company technical expertise [2].
Some postgraduate studentships also involve direct links with industry, e.g. the CASE (Cooperative Awards in Science & Engineering) studentships offered by the UK Engineering and Physical Sciences Research Council, for which the student has an industrial sponsor who provides both a top-up grant and a degree of supervision.
Individual members of staff in a university may act as consultants to industry and commerce, thus providing a route whereby their own expertise can be exploited commercially. This may generate income for the individuals concerned, but universities vary in the extent to which they attempt, or indeed are able, to ensure that a share of this income is returned to the institution.
Patenting of research ideas is a seemingly obvious route to commercialisation. However, patenting is essentially a mechanism for protecting intellectual property rights, and not necessarily a way of exploiting them. Furthermore, patenting is to some extent a cultural phenomenon. The complaint raised in Hong Kong for instance [1], about the small number of patents taken out by the five Hong Kong research universities, was made in relation to the performance of American universities such as MIT. Patenting is almost a way of life in the USA, along with the ensuing litigation involved in trying to protect the patents, and does not necessarily improve the commercialisation process. In the case of IT, patenting of software is in any case notoriously difficult.
Commercial companies which sell software effectively issue a licence-to-use with each copy of the software, but have to go to enormous lengths to catch up with those who pirate their software and sell it for their own, and not the originator's, profit. Universities clearly cannot get into the situation in which these companies find themselves, and need to proceed on a different basis.
The options are for the originator(s) of the software to form a spin-off company, for the software to be sold or licensed to an existing commercial organisation, or for the software to be distributed free of charge, usually to other academic institutions. In the latter case there is still some virtue in using a licensing mechanism. This ensures that the software remains consistent at all sites, and that any improvements which individual users might make can be made available to all users.
This mechanism has been used effectively in the case of the ML programming language, for example, created in the Laboratory for Foundations of Computer Science (LFCS) at Edinburgh University. The licence states that "The licensee may modify the supplied ML System on condition that any significant changes are notified to LFCS and made available to LFCS such that they may be incorporated within future releases of the ML System and licensed under the conditions of this licence agreement.".
Of course some would argue that it is not the function of universities to produce industrial quality software, but James Larus, author of the very succesful, in terms of its use in many universities, simulator for the MIPS instruction set (SPIM), has argued [4] that writing `real' software in universities is important both from the point of view of embodying new research ideas and of producing pedagogical tools (such as SPIM). It also ensures that those who teach will also have proved that they can do.
Larus believes that academics should make their software freely available to other academics (for fame), but should charge a nominal fee ($1000) to industry for evaluation or research use, and that universities corporately should negotiate terms for serious commercial use of university originated software. (There is an implication here that universities should be clear about their own internal arrangements for the ownership of software - is it the originator, the university or both?). In practice university software may only be perceived to have any value once it has an established base of users (c.f. the Unix operating system, for example).
Nevertheless, the University of Edinburgh, for example, has spawned a number of successful IT spin-off companies. Amongst these are Clan Systems, Edinburgh Portable Compilers, Lattice Logic, 3L, Memory Corporation, Vision Group and Wolfson Microelectronics. The University now actively encourages this process through its Technology Transfer Centre (TTC), set up in 1988. TTC is a joint venture between the University and the City of Edinburgh, intended to act as an incubator for small start-up companies.
One of Edinburgh's earliest spin-off companies was Lattice Logic, which was formed in 1980 when a lecturer in Computer Science, John Gray, left the University with two PhD graduates. Lattice Logic produced a novel Computer Aided Design system for the automatic layout of integrated circuits in the gate array style. This system took a description of the desired structure of a chip design written in an innovative hardware description language (Model) and, following simulation and verification, generated the mask-level fabrication data. The company supplied its system to a number of high profile enterprises including Ferranti Ltd and the pan European company ES2. ES2 eventually bought Lattice Logic in the late 1980s and many of the software team, though without the founders, still work together, now as part of the Cadence Design Systems company's offices in Edinburgh.
The Vision Group (originally VLSI Vision Ltd) was started in 1990 by Professor Peter Denyer of the Department of Electrical Engineering at the University of Edinburgh. This spin-off company was set up in the light of significant commercial interest in research which Denyer had carried out with Dr David Renshaw from 1985 into CMOS image sensors and mixed signal sensor-processors. The research culminated in demonstrating the world's first single chip video camera and for which IEEE best paper and best circuit awards were received.
Vision Group began life in the University's Technology Transfer Centre and was initially financed by venture capital arranged by Unived Technologies Ltd, the University's industrial liaison company. In 1992 VVL outgrew the TTC facilities and re-formed using privately raised capital as an independent company. Since then VVL has continued to grow its business, designing, manufacturing and trading in video camera technology and image processing systems.
In the early days the University did not seek to reap any direct reward for its contribution to the intellectual capital of these companies. More recently it has taken a more proactive stance, mainly through Unived. As a means of reaping its own reward for having provided the environment in which the Vision Group research took place, for example, the University took shares in the company when it was set up. These shares are now being sold at a significant profit and the funds re-invested in equipment and postgraduate studentships in the Electrical Engineering Department and in the University's estate.
The vision was timely, since a number of national agencies (the Computer Board for Universities, the Department of Trade and Industry and the (then) Science and Engineering Research Council) were also becoming conscious of the need to develop expertise in parallel computing and were able to make funding available for this purpose. EPCC was successful in bidding for a significant proportion of this funding, and has continued to be successful in obtaining other funding both from the UK and from the European Union. It is now well established as the leading, cross-disciplinary European centre applying High Performance Computing (HPC) to real-world problems, and for the past eight years has successfully commercialised scientific research. EPCC has an active and diverse portfolio: industrial consultancy, research projects, world-class facilities, and an unrivalled range of training programmes [5].
As a technology transfer centre, EPCC's role is to provide expertise and services to industry and commerce in computational modelling, data analysis and other areas. It has developed a broad skills base which includes HPC techniques and computer science, physical sciences, engineering and earth sciences. Its services include consultancy, adaption of applications, development of solutions, implementation of international standards (e.g. the second Message Passing Interface (MPI-2) standard), and support for production use of its extensive HPC systems. These systems include the largest (512-processor) Cray T3D in Europe (and the most heavily used in the world) and a 240 processor Cray T3E system.
Despite its reputation and proven competence, finance and commerical organisations do not perceive EPCC, as a part of a university, as being an appropriate supplier of packaged software, either in terms of being able to provide on-going support or of underwriting any potential risks. Thus in order to commercialise some of the software developed in EPCC, a number of its members left in 1995 to form a spin-off company, Quadstone, which supplies data mining and traffic simulation software to a number of clients worldwide.
The Centre's relationship with the University is synergistic. EPCC is self-financing, through income generated by the provision of services to industry and academia, but nevertheless derives great benefit from being part of the University, which provides it with a rich mixture of scientific stimulus and logistical support. In return EPCC contributes to the research standing of the University and in addition generates a significant income stream by payment of overheads earned from its research and development contracts.