Computing Profession Viewpoint

(Mis)Managing Parallel Computing Research Through EU Project Funding

Considering the possible negative long-term effects of EU projects on European scientific culture.
  1. Introduction
  2. EU Projects: "Novel Ideas for Radically New Technologies"
  3. Not-So Novel Ideas for Radically Better Funding
  4. References
  5. Author
  6. Footnotes
(Mis)Managing Parallel Computing Research through EU Project Funding, illustrative photo

Much high-performance and parallel computing research in Europe is funded by the European Union through the so-called "EU projects." These projects are highly competitive and often considered highly attractive and prestigious because of their scientific quality, involvement of outstanding researchers, contributions to the education of new Ph.D.’s for academia and the European workforce, and opportunities for collaboration with industry in Europe.

I will argue a dissenting opinion that EU projects in this field may not serve their stated purposes of advancing science and education, and are not effective in subsidizing European industry. On the contrary, they are a time-consuming effort with often mediocre outcome that tend to perpetrate a narrow, self-referential, and trivialized view of the scientific process with possibly detrimental long-term effects on the scientific culture in Europe and the quality of the workforce. The problems are not exclusive to the EU funding system, and also not to parallel computing research, but to a large extent caused by excessive managerial overheads and the specific stated requirements and unstated expectations in the EU project conception. This Viewpoint from the perspective of an academic researcher is based on long, personal experience with EU projects in parallel computing and related areas, in different roles, and with different degrees of "success." Since these reflections are not meant as criticism of specific projects or activities, I will not cite concrete project figures or outcomes, but contend that my claims can be backed up by publicly available material.

Back to Top

EU Projects: "Novel Ideas for Radically New Technologies"

EU projects are set up in response to regular calls by the European Commission that outline the issues to be investigated and the kind of projects that may receive what kind of funding under which conditions. The calls ask for novelty, excellence, radicalism and breakthroughsa and also outline the composition of eligible consortia that will consist of "leading experts" from European university research groups and (often, but not in all call types) European industry or business, preferably small and medium-sized enterprises (so-called SMEs). Funded projects typically run for three years, and are expected to contribute to education by producing excellent, new Ph.D.’s for the European workforce. Exploitation and dissemination are of the utmost importance to demonstrate uptake and impact. Progress is documented in extensive "deliverables" and "prototypes," and constantly monitored and assessed by EU "experts." Deviation from work plan is an exception and needs to be negotiated. In not-so-large EU projects, five to 10 groups can be involved, each with a number of people that meet as a whole or in smaller groups several times a year. As part of the dissemination activities it is important for all consortium members to show regular presence at various associated, often self-initiated workshops and conferences, including meetings and workshops of other projects; high publication output is encouraged.

There are several points to stress. First, research areas are determined extrinsically by the European Commission (in consultation with "leading experts") which is in itself not an intellectually contributing peer of the scientific (parallel computing) community and therefore only in an indirect, not independent position to define scientific research agenda. Second, projects are to be tightly managed, documented, and overseen by the appointed Commission experts. Third, project proposals themselves, which after approval turn into detailed "Descriptions Of Work" (DOW: work-packages, which deliverables when, "milestones," month-by-month work, exploitation and dissemination plans), are themselves considerable pieces of work (easily 50 or more pages) taking weeks of intense effort to prepare. Fourth, dissemination of results and industrial exploitation are key factors in successful proposals and projects.

It is doubtful whether truly groundbreaking research can be planned as envisioned by the Commission managers and thrive under the EU project conditions.

Now, it is doubtful whether truly groundbreaking research can be planned as envisaged by the Commission managers and thrive under the EU project conditions; but "normal science" is rarely called for. The effort in proposal writing is wasted (unless recycled) if a project is not funded, which lead proposals to promise what is called for, namely a meticulously documented, steady flow of high-quality breakthroughs that can readily be taken up by industry. And although projects can have large, total volume in terms of money and personnel, the managerial overhead and enforced geographical and thematic spread among consortium members in reality means that time and personnel per group is often too small to pursue serious and realistic research. Fortunately, since projects define their own "verifiable success criteria" and "key performance indicators," success can nevertheless be guaranteed, either because results to be produced were safely known in advance, and therefore per definition not groundbreaking, or because goals, despite the inflated language, are set so low or made so vague that failure is impossible by design, as long as deliverables are delivered, milestones met, prototypes produced, workshops and dissemination activities organized, and consortium management business attended to. Whether proposers actually believe what they propose, regarding both goals and execution, is an open question. However, proposal-writing skill is certainly cultivated by the EU system, and those in command of the language needed for successful proposals can have a welcome role as partners, or consultants in the proposal phase.

Since running an EU project and preparing for the next one takes time, the leading experts who formulated the project in the first place may not be factually contributing to the project in any major way, nor be supervising inexperienced junior personnel carrying out the projected work. Instead, young researchers in need of guidance for their Ph.D. may find themselves left alone with the considerable responsibility of advancing the project from which they are paid, including managerial duties and the writing of deliverables whose sole merit and purpose is to fulfill what was promised in the DOW. From a scientific point of view, EU project deliverables are often of a sub-par standard with little of value to the student or to the scientific community. Since the project must be successful, there is severe temptation to willingly or innocently sell trivialities as breakthroughs, effectively installing and reinforcing a belief in the student and the interested public (such as via newspapers) that trivialities are major breakthroughs. The same can be observed for the prototypes that are important as "proofs-of-concept" and means to "transfer knowledge" from research to industry, which are often of low quality and designed only to survive a single, superficial Commission expert review. Thus, EU projects may seem not to contribute to develop scientific attitude, conduct and practice in the Ph.D. student. A more experienced postdoctoral researcher may likewise be left to herself with the day-to-day running of EU projects without having enough time to concentrate on the research and lasting results that will be needed to eventually secure a permanent position.

The role of the industrial partner in an EU project is sometimes that of a peer, sometimes that of an advisor, sometimes of a provider (in the case of high-performance computing: of access to technologies), but often that of a politically necessary appendix. Especially SMEs can be as dependent on the EU funding as small university groups, but may not have the capabilities and resources to contribute significantly to research-oriented deliverables, and often do not, thereby leaving more of the project burden on the academic partners. On the other hand, SMEs sometimes expect concrete product development to be carried out in the project for which the academic partners typically have neither the skill nor the manpower. Peculiarly, large (non-European) multinational companies can receive subsidy as EU project partners. It is not always obvious that this contributes to the projects, neither why these companies should be recipients of EU funding.

Exploitation and dissemination being important factors in EU projects lead to a wealth of activities serving to document successful "take-up": publications, workshops and meetings at conferences, interproject workshops, user training, contributions to standardization bodies, and so forth. The primary purpose of many of these activities is to meet the dissemination plans, and has led to a proliferation of workshops presenting and publishing project-relevant trivialities. Apart from the time this consumes, the apparent authority of a workshop masquerading as a scientific event at a well-established conference may reinforce docility and low standards in the naive Ph.D., and appall and deter the more observant one. As harmless as many of these time-consuming activities may be, premature or incompetent standardization is not, and can bind considerable resources to either comply with or fight a bad standard, thus hampering both scientific and industrial progress.

Back to Top

Not-So Novel Ideas for Radically Better Funding

European research groups in high-performance and parallel computing are materially dependent on external project funding, and EU projects give a comfortable headroom with the attractive possibility to collaborate closely with other European research groups. However, the drain by the management of projects and the effort required for the overly detailed project proposals, are stunting. It would be liberating and productive to get rid of both.

First, since there is no proof that strict management procedures actually lead to better and more lasting scientific outcome, why not drop the Commission involvement entirely from the project execution (as in some national schemes), and trust responsible projects to select adequate means for managing themselves, tracking progress by short summaries and actual outcome instead of excessive technical deliverables? Likewise, the overemphasis on dissemination and take-up activities should be moderated; well-run projects are themselves capable of deciding when and what makes sense.

It would be more transparent and less ambiguous if industrial participation in EU projects was not directly funded through the projects.

Second, it is an equally unproven assumption that scientific breakthroughs can be predicted by the promises made in project proposals. It would save much painful effort and dear time to ask instead for short, simple, matter-of-factly, problem-oriented funding applications without inflated promises, without excessively detailed work plans and without politically motivated constraints on consortium composition. With clear, and not too restrictive criteria for funding eligibility that can realistically be checked by the available Commission experts, it would perhaps even be acceptable to acknowledge the fact—given the current oversubscription and low success rates—that funding decisions among eligible proposals are … random. Criteria could emphasize previous scientific results of the involved researchers, along the lines sketched by Thorup,1 and in general focus more on actual instead of projected outcome.2 This would imply that proposals and projects to a larger extent be driven by individual groups with a clear vision and consortium members selected by their specific project-relevant expertise. It would make long-term scientific, and perhaps even commercial sense to make it possible and easy to extend projects that produce actual, great results or valuable prototypes; currently, such extensions are difficult and often depending on a fitting, next call. Since the three years that EU projects typically run is in many cases too short to complete a Ph.D. thesis and gain the necessary breadth and depth beyond the horizon of the subsidizing EU project, means of after-project follow-up funding are especially important for Ph.D. students; the practice of putting the student on whatever next project may be available is hazardous.

EU projects in high-performance and parallel computing sit between many chairs, and in many cases industrial involvement makes a lot of sense. However, it seems confusing at best to (ab)use scientific research projects to subsidize European (SME) industry. Can’t this be done much more effectively by direct means? In any case, it would be more transparent and less ambiguous if industrial participation in EU projects was not directly funded through the projects. Strong projects would be interesting enough that industry would want to participate out of its own volition, which in particular should be the case for large businesses.

Finally, it seems that the role of European-level funding should be to support what cannot be funded nationally, which on the other hand implies that strong national funding of research groups and projects is still needed. In addition to existing, large-scale European high-performance computing infrastructure with personnel for support and development, more possibilities for travel grants and lightweight working groups to foster contacts between European research groups, and more EU scholarships for Ph.D. students doing their Ph.D. at different European universities would also be welcome additions.

Back to Top

Back to Top

Back to Top

    1. Thorup, M. Funding successful research. Commun. ACM 56, 3 (Mar. 2013), 38–39.

    2. Vardi, M.Y. Incentivizing quality and impact in computing research. Commun. ACM 58, 5 (May 2015), 5.

    a. Bertrand Meyer addressed this paradigm-shift mania and its effects on scientifi attitude in his "Long Live Incremental Research" blog posting, see http://bit.ly/2dFOV75; prose from recent FET Open calls shows how up to date this still is.

Join the Discussion (0)

Become a Member or Sign In to Post a Comment

The Latest from CACM

Shape the Future of Computing

ACM encourages its members to take a direct hand in shaping the future of the association. There are more ways than ever to get involved.

Get Involved

Communications of the ACM (CACM) is now a fully Open Access publication.

By opening CACM to the world, we hope to increase engagement among the broader computer science community and encourage non-members to discover the rich resources ACM has to offer.

Learn More