Select Committee on Science and Technology Second Report


CHAPTER 7: UK INDUSTRY AND THE RESEARCH FRAMEWORK

Introduction

7.1  This Chapter first sketches out the major transformation of the UK computer industry over the last two decades. New and world-leading strengths have emerged, although these are not generally recognised. The Chapter concludes by considering the present framework for research funding.

The UK computer industry

7.2  As noted in previous Chapters, the United Kingdom had a prominent role in the development of computers. Several early computer manufacturing companies formed, which, through a series of take-overs and mergers, led to the formation of ICL as the single UK mainframe manufacturer.

7.3  ICL enjoyed a period of business success but, with the rise of the PC and the growing dominance of the microprocessor as the engine of all computing, the international mainframe industry went into decline. ICL was bought by Fujitsu, withdrew from leading-edge manufacture, and now operates as a computer services company. Recently, the ICL name has been dropped, and this has been seen by many as the end of UK involvement in the computer industry.

7.4  However, just as the UK's hardware manufacturing business was declining, a new wave of companies was forming. These companies were based in the ascendant microprocessor technology. The first — Inmos[61] — envisaged the development of large parallel computing systems based on arrays of single chip computing elements or "transputers". Although the technology was successfully realised, the radical computing model gained only limited market acceptance[62]. The company was subsequently bought by ST Microelectronics and its technology survives today in the form of the ST20 processor core, which is used in a number of embedded applications. The Inmos experience left a rich legacy of experienced microprocessor architects around whom, as noted by the National Microelectronics Institute (p 209), several new architecture-based companies have formed.

7.5  Also in the early 1980s, Acorn Computers Limited was developing the BBC microcomputer (which resulted in the first computer-literate generation of UK school pupils). To enhance their educational computing platforms, they then developed the 32-bit ARM microprocessor. This was much less radical in concept than the transputer and, serendipitously, was ideally suited to the low-power mobile computing platforms of the 1990s.

7.6  Acorn's desktop computing platforms were overtaken by the now ubiquitous PC, but the microprocessor technology was spun out into a separate company — ARM Limited — and thrived. It has become the flagship of the UK chip design industry and leads the world in the design of simple, low-power 32-bit microprocessors for highly-integrated SoCs.

7.7  Around these core microprocessor companies, others have sprung up. Professor May commented that "We have a new version of a computer industry, probably quite an important one, growing in our midst, and we have not noticed yet" (Q 253). From the developments outlined above, a UK computer industry is emerging which is quite different in nature from its antecedent. Computing is already pervasive, and is clearly set to become more so. The industry is thus of great importance as well as economic significance.

7.8  To enable the UK's new-style computer industry to be appropriately recognised and fostered, we recommend that the Government urgently establish a task force to bring greater coherence to their dealings with the industry, through the implementation of the recommendations in our Report and otherwise. Further, we recommend that the Government give wide publicity to the composition, remit and contact arrangements for this task force.

7.9  A priority job for the new task force will be to ensure that, as a foundation for proper development of policy, information collection and dissemination at all levels of Government (including Regional Development Agencies) properly reflect the new-style UK computer industry.

The UK academic research environment

7.10  UK universities engage in a range of relevant research. Research into semiconductor manufacturing processes and alternative device technologies usually takes place within electrical engineering, physics and material science departments; research into design and architecture of microprocessor systems is normally within the remit of electrical engineering and computer science departments.

7.11  The university funding regime imposes financial constraints on the scale and nature of university research activities. In many universities, academic staff numbers are limited by the demand for undergraduate places in those subject areas. Demand has been dwindling for places on electrical engineering courses, leading to shrinking staff complements[63] and, in some cases, the closure of departments. Demand for places on computer science courses is buoyant, though the majority of applicants are seeking courses in IT rather than in subjects relevant to the design and manufacture of microprocessors.

Funding for UK academic research

The dual-support model

7.12  The present funding arrangements for UK universities are based upon two principal sources of funding. Funding Councils — in England, the Higher Education Funding Council for England (HEFCE) — provide baseline funding for teaching, research and infrastructure. The Research Councils then provide funding for the additional costs of specific research projects.

7.13  As HEFCE noted (p 199), research funding is linked, on a subject-by-subject basis, to research ratings that are established every five years through a national Research Assessment Exercise (RAE). Each subject area in every UK university is given a rating from 1 (the lowest) to 5* (the highest) on the basis of submitted evidence of research quality. The higher ratings (particularly 4, 5 and 5*) result in an increased baseline funding that enables the recipient department to operate at a more favourable student/staff ratio, giving staff a lower average teaching load and more time to engage in research.

7.14  Overall, however, the total funding available to universities falls well behind that which would be required to fund staff salaries and university building programmes at a rate competitive with comparable groups in industry or universities in other countries. We will consider staffing further in the Chapter 9, but note here that UK facilities for academic research are suffering as a result of long-term under-investment. Dr Taylor commented (Q 22):

"We need to increase the volume of research funded in the United Kingdom by at least 7% per annum above inflation for the foreseeable future so that we can begin to catch up with the levels of spending in our major competitor countries. In this current spending review we are faced with under-funding, over-trading, infrastructure issues, salary issues and the need to continue to increase the volume in real terms of the work we do."

Dr Keddie added (Q 22):

"Looking at it from a business science point of view, there is absolutely no doubt that unless we invest sufficiently in the science base we will not be generating knowledge fast enough for the business/commercial side of things."

EPSRC

7.15  The main Research Council funding for UK academic research into semiconductor process technologies and microprocessor design and architecture is provided by the Engineering and Physical Sciences Research Council (EPSRC). It receives an annual budget of £450 million from government which it allocates to a number of research and postgraduate training programmes, with £300 million going into research project grants. The areas of primary interest to this Inquiry are covered by the Information Technology and Computer Science (IT&CS) programme[64], with further additional investment in the underpinning technologies coming from programmes such as the Materials programme. The IT&CS programme has an annual budget of £67.5 million and, as noted by Dr Taylor (Q 29), the total annual funding including materials is around £120 million. The overall balance is more towards work on materials and devices than it is towards design and architecture.

7.16  The majority of EPSRC funding is allocated through responsive mode programmes where there is no constraint on or direction given to the area of research. A research proposal is refereed anonymously by relevant experts, and the referees' reports are then assessed by an expert panel which prioritises a batch of proposals. The available funds are divided between different areas by negotiation within EPSRC, and the projects that get funded are those at the highest priority that fall within the available funding. The success rate can be very low — Dr Cowburn noted (p 80) a recent Physics panel where only 16% of proposals were funded.

7.17  The other main EPSRC funding stream is through managed programmes. Here, EPSRC identifies a particular priority area for funding through its Technical Opportunities Panel and the User Panel (Q 48) and invites proposals in that area. As Professor O'Reilly noted, the historic trend in EPSRC funding has been away from managed programmes, securing greater vibrancy from an increased proportion of responsive mode funding (QQ 49 & 50). The Minister indicated that EPSRC was now moving somewhat back from that, being readier to build up capacity in particular areas (Q 479).

7.18  In addition to these two basic modes of funding individual research projects, EPSRC also has mechanisms for encouraging particular styles of research such as IRCs (Interdisciplinary Research Collaborations), the Adventure Fund (a new initiative to encourage more ambitious blue sky projects), visiting research fellowships, and so on.

7.19  A notable and particularly relevant IRC has been established in quantum information processing[65]. EPSRC announced in January 2002 £9 million funding for this over 5 years. This is a significant development in UK support for the radical technology of quantum computing[66].

e-Science

7.20  The Research Councils' £98 million e-Science programme is an activity across the various Research Councils to develop the technology and applications that will enable large-scale science to be carried out through distributed global collaborations enabled by the internet.

"Such collaborative scientific enterprises will typically require access to very large data collections, very large scale computing resources and high performance visualisation back to the individual user scientists."[67]

e-Science is the application of Grid technology (see paragraph 6.11) to support scientific collaboration.

PPARC

7.21  While EPSRC is the major source of UK government funding for research into semiconductor device technologies and microprocessor design and architecture, there is also some funding in this area from the Particle Physics and Astronomy Research Council (PPARC). Particle physics experiments require very high levels of computation to process their results, and the forecast computational requirements for planned experiments suggest that they will exceed the capabilities of available machines in the near future. PPARC has therefore initiated a major activity to develop an application-specific computing capability to address this requirement (p 24).

7.22  While we note the very demanding computing requirements of PPARC's research base, we recognise that the history of special-purpose computer hardware is not encouraging. The economic momentum behind advances in general-purpose computing tends to make any gap between the capabilities of general and special-purpose hardware short-lived.

Government laboratories

7.23  Government Laboratories such as NPL (see paragraph 4.40) carry out research in specific domains with direct Government funding. Some such laboratories have been privatised over recent years, but there remain some national facilities maintained because of their strategic contribution to the nation's economy, standing or well-being.

Interdisciplinary work

7.24  Interdisciplinary work (in universities and elsewhere) is essential for advances in the areas of interest to this Inquiry. We note that there are several real or perceived obstacles to such work. Severe internal competition for resources both between and within universities may incline departments to retreat to traditional areas where they believe they can do well and may disincline them from exploring novel areas at the interface with other disciplines. Furthermore, internal university processes for resource allocation and cost attribution may make interdepartmental initiatives awkward.

7.25  Computing technology makes demands on the skills found not only in computer science departments but also in departments for electrical engineering, physics and mathematics. Future developments are likely to involve ever more diverse departments, for example in the biological sciences. We recommend that those universities which wish to be at the forefront of strengthening the United Kingdom's stake in computing ensure that their structures and processes encourage interdisciplinary research collaboration between departments.

7.26  Such difficulties are exacerbated by the belief, widely held within universities, that interdisciplinary work tends to be harshly graded in the Research Assessment Exercise by experts in the mother fields. The Funding Councils have in the past declared that this is not the case but the perception and its damaging consequences remain. In response to a Report by the House of Commons Science and Technology Committee[68], HEFCE have undertaken to examine the problem as part of a wider review of the RAE.

7.27  We applaud the initiatives taken by the Research Councils through the programme for IRCs and believe that, although the earlier tranche of Interdisciplinary Research Centres may not have fully achieved the Councils' original objectives, there were beneficial consequences for interdisciplinary working with universities. We therefore recommend that the Funding Councils and Research Councils take further positive steps to ensure that interdisciplinary work in the field of microprocessing techniques is not only actively encouraged but also manifestly assessed in an even-handed manner.

7.28  One of the concerns that led to our Inquiry was that EPSRC had not been quick to recognise and exploit the synergies that may arise between different parts of its responsive mode programme. Although this is not a matter that we looked at closely, it clearly is something that should be done by a pro-active and innovative research council. Indeed, the DTI and Research Councils should also be actively looking for ways in which elements of the sometimes over-compartmentalised science base might usefully be brought closer together in pursuit of interdisciplinary objectives.

EU funding

7.29  In the 1990s, a major focus of EU funding in the microprocessor area was the Open Microprocessor systems Initiative (OMI), aimed at encouraging the widespread take-up of SoC technology by European industry. One major UK beneficiary of this funding programme was ARM Limited, whose evolution from start-up to world-leading player substantially overlapped the period of OMI funding. As noted by Commissioner Liikanen (Q 517), the new EU framework programme for information and communication technology is based on developing the concept of ambient intelligence — the idea that computing will become all-pervasive.

7.30  There is much anecdotal evidence that UK industry is somewhat reluctant to participate in EU programmes. This was reinforced by EU officials who noted an interesting difference between the United Kingdom's and other Member States' financial participation in relevant European programmes (Q 563). Although the UK's overall participation is in line with comparable Member States (such as France and Germany), the distribution of the funding between the various sectors is noticeably different, as below.

  
EU average
United Kingdom
Universities
24%
52%
Research centres
19%
4%
SMEs
24%
19%
Large companies
28%
21%
Public organisations
5%
4%
Total
100%
100%


7.31  Although there are difficulties in interpreting the EU's figures, they are consistent with the general view that there are opportunities for UK industry to do better in accessing European programmes, with the potential dividend of capitalising on some valuable opportunities in the international computing business. Accordingly, we recommend that the DTI should investigate the reasons for and the implications of the apparently poor participation of UK industry in EU programmes (particularly those relevant to microprocessing) and, as necessary in the light of the findings, to take action to encourage greater participation.

Industrial research funding

7.32  Research does not, of course, have to be publicly-funded. The private sector can have a significant role. Industrially-funded research may be carried out in universities or within industrial research laboratories. We were pleased to note that the CBI has produced a guide on business working with universities: Partnerships for Research and Innovation between Industry and Universities: a guide to better practice[69] (Q 444).

7.33  However, a DTI survey into UK levels of industrial research funding showed that, except in the pharmaceuticals industry, the levels of research were between a third and a half of the levels in competitor countries (QQ 163 & 450). As the IoP noted (QQ 165 & 172), reasons for this disparity may be that UK industry management is risk averse; financial analysts do not view R&D as an asset; and, perhaps most significantly, any return on the investment is beyond an unreasonably short time-horizon of as little as three or four years.

7.34  Pay back times elsewhere in Europe can be at least twice that and, as noted by the BCS (Q 131), US companies take an even longer term view. That was borne out during our visit to Silicon Valley when we heard that IBM sought a return on basic research over a period of 10 to 15 years.

7.35  There is, as the IoP noted (Q 172), no penalty for missing opportunities — at least in the short term. The penalty not only comes much later but can also be obscured by general company decline. Noting that CBI members found the UK environment for industrially-funded research in academia generally more favourable than that in the US (Q 445), we recommend that UK companies and the finance sector should pay more attention to investing in R&D in the field of microprocessing. This will enable companies to maintain and develop their markets in the longer term.

Aggregate spending on R&D

7.36  In its written evidence (p 1), DTI supplied a useful international comparison of aggregate spending on relevant R&D. Not least for its final sentence, this is worth repeating in full.

"Major European countries spend significant sums on R&D for information and communications technologies (ICT). The 2001 OECD Science, Technology and Industry Scoreboard[70] shows the major European countries spends as:

  
US$ billions
Germany
5.7
France
3.8
United Kingdom
2.2
Italy
1.8


"This is such that European spending taken together is comparable with that of Japan — at $22.2 bn (in fact, the EU15 slightly exceeds Japan) but the US spends roughly twice this amount ($46.62 bn). The biggest spender on ICT R&D in Europe in 1999 was Sweden (mainly Ericsson) with Finland being the biggest spender in GDP terms (mainly Nokia). However, to put these figures in context, Intel has maintained its R&D spending throughout the recent downturn in the industry at just under $3 bn per annum."


61   The commercial and political factors surrounding the setting up of Inmos through to the production of the first transputer chips in 1985 is described in The Inmos Saga - a triumph of national enterprise? by Mick McLean and Tom Rowland, pub. Frances Pinter (London), 1985. ISBN 0-86187-559-1 Back

62   In part because of the difficulty (as noted in paragraph 5.8) of developing software for general-purpose parallel processing.  Back

63   Even so, as discussed in Chapter 8, there is a crisis in recruiting such staff. Back

64   During the course of 2002 the IT&CS programme has been renamed ICT (Information and Communication Technologies). Back

65   www.materials.ox.ac.uk/qipirc/ Back

66   See paragraph 4.33(e) and Box 8. Back

67   www.research-councils.ac.uk/escience/ Back

68   The Research Assessment Exercise: Government Response, Fifth Special Report Session 2001-02, HC Paper 995 - the response to the Committee's Second Report Session 2001-02, HC Paper 507. Back

69   CBI, April 2001 ISBN 0 85201 553 6 Back

70   See www.oecd.org/EN/document/0,,EN-document-54-1-no-15-17270-54,00.html Back


 
previous page contents next page

House of Lords home page Parliament home page House of Commons home page search page enquiries index

© Parliamentary copyright 2002