Open road to OMI sweet home
Julian Newman looks at the new face of European research and finds hope for the future direction of funding
The European Union has recently launched its fourth Framework programme for Research and Technology Development (RTD). To many in the IT research community, Esprit and RACE are synonymous with European funding. These names will vanish, and all our funding hopes will be pinned to Framework 4.
Most previous schemes were nominally part of Frameworks 1 to 3, but were generally perceived as discrete entities, and varied considerably in aims, procedures and modus operandi.
Is this a real change, or merely a change of name? Some parameters of European funding will be unchanged. Our old friend "subsidiarity" will ensure that there will be no support for a project based in a single state, no matter how meritorious its scientific content.
United Kingdom applicants will no doubt continue to be "over-represented" in Framework 4 as they were in Esprit, a fact that partly reflects the usefulness to a consortium of having native English-speakers to write their proposal documents.
But this also highlights how the lack of national funding forces British researchers to look to Brussels: paradoxically a Euro-sceptical Government is acting in the best possible way to create a Euro-federalist scientific community.
But one way in which Framework 4 will differ from its predecessors should undoubtedly please the British government, perhaps to the dismay of academics. With the Gatt negotiations well behind it, the union need no longer be quite so sensitive about the fine distinction between developing the technology base (fair) and industrial subsidies (foul). This will intensify a trend apparent in the later years of ESPRIT: to a greater extent than before, European RTD projects will be "industrially led", and academic partners will be judged as much on their capability to contribute towards industrial goals as on their scientific excellence.
While remaining "pre-competitive", projects will also be nearer to the market: MEPs have increasingly wanted proof that RTD funding leads to concrete benefits in a form they can understand; the most obvious way this can be achieved is when project results are embodied in a marketable product.
Framework 4 funding for information technology was secured despite lobbying by the biotechnology industry, which claimed that Esprit had failed to save European IT: further funding would simply be the reinforcement of failure -- the money should go to biotechnology instead.
The exploitation plans of the industrial partners, already an important part of proposals, will have a central place in the evaluation process.
Proposals will continue to be evaluated by independent experts, to ensure an independent judgment of quality. However, there are subtle differences from peer review as practised by research councils.
First, the criteria which experts are asked to use -- and which are clearly stated in the calls for proposals -- reflect the applied, pre-competitive but market-oriented nature of the programmes.
Second, there is an increasing trend towards judging projects not independently, but as a contribution to a portfolio of European RTD designed to meet definite strategic objectives. The clearest example of this in Framework 4 is the use of a number of Focussed Clusters of projects.
To see how Focussed Clusters may work, we have one example. The Open Microprocessor systems Initiative (OMI) which originated under Esprit but will be carried forward as a Focussed Cluster under Framework 4.
OMI is not only about microprocessors: it is also intended to extend the philosophy of "open systems" to the microprocessor level. The impetus to do so came from a recognition that technological changes -- in particular the possibility of creating full-scale computers on a chip -- constituted an opportunity and a threat.
The opportunity is, for example, to create new heterogeneous systems (combining processors with different architectures) to satisfy the requirements of specific applications.
So instead of the application-specific integrated circuits we have now, we could have application-specific integrated processors. The most sophisticated of these might use operating systems originally developed for workstations and general-purpose computers; so a high level of intelligence could be embedded in telephones, printers, automotive systems and so on.
The threat is that as more and more "disappears into the silicon" system integrators may have less scope for differentiating products and be tied more closely to particular suppliers: the seriousness of this threat is apparent when it is considered that the European electronics industry is 80 per cent dependent on imported components, as against 16 per cent in the United States and 33 per cent in Japan.
Europe's position in important aspects of software is weak: we are good at developing multimedia applications, for example, but the tools we use to develop these applications are most likely to have come from the US.
There is virtually no such thing as a European programming language compiler industry at all.
We still retain a significant capability in operating systems and distributed systems. But the destruction of the native systems integration industry would still further impair Europe's position in software. There are undoubtedly benefits ("externalities") for a firm in being close physically to the source of innovation in the technologies on which it depends with corresponding penalties if the source of innovation drifts away.
OMI is an attempt to counter these threats. Therefore, it addresses software as well as hardware, and is driven by user needs, technical opportunities and a recognition of the need for open standards. There has been significant user involvement in project definition and in applications studies.
For example, the ARCHIE project, in which Dundee University is a partner, involves the development of multimodal interfaces -- of particular relevance to the disabled, but also with applications in aerospace -- the major interest of the industrial partners. Such applications studies are not directly concerned with microprocessors, but help to identify user requirements which the OMI library of microprocessor "macrocells", known as ELI, should serve to support.
There were extensive industrial consultations before the launch of OMI, as a result of which a strategically-required set of projects was identified; this strategy was built into the workplan against which consortia bid in the normal competitive process.
Successful bids were not confined to consortia whose members had taken part in the consultative process; nor was the fact that a project had been identified as a strategic requirement within the OMI any guarantee that at least one project conforming to it would be successful. Some of the strategic requirements proved difficult to meet; but it is a notable characteristic of OMI that when such a situation arises, the OMI community and the Commission work together to help ensure that a successful resolution of the problem can eventually be achieved.
Before OMI, Esprit was fragmented into microelectronics, information processing systems and software, computer integrated manufacturing and advanced business systems and peripherals.
Particular OMI projects are located in the most relevant Esprit technical area; but OMI cuts across these divisions, and in this sense is a highly "political" initiative. Any researchers involved in new Focussed Clusters, such as Technology for Business Processes, may well experience a certain degree of "creative tension" arising from this strategically-driven breaking of established boundaries.
The essential lesson for academics wishing to be involved in the Focussed Clusters, is that they need to establish early and organic relationships with firms who have a genuine need to exploit anticipated results.
Marriages of convenience between academics and firms will not lead to high-quality proposals. Exploitation plans which are vague and unspecific will not cut ice in the current climate. Focussed Clusters should not be seen as cosmetic packaging, but as an attempt to ensure that Framework 4 achieves clearly-articulated economically significant strategic objectives.
Julian Newman is reader in business computing at Glasgow Caledonian University.