Gridlocked by transfer

The lack of institutional lines of command can stall collaborative projects, says Peter Coveney.

Academics today find themselves under considerable pressure from a number of different demands, including research, teaching and administration. The problems of dealing with all these demands are compounded by the difficulties we have in getting professional institutional support for our endeavours despite the substantial slice that universities receive by way of money for indirect costs ("overheads") on grant income.

Since the end of 2001, I have been trying to run RealityGrid, a £3.4 million-plus high-profile e-science pilot project funded by the Engineering and Physical Sciences Research Council. RealityGrid has seven academic partners and as many non-academic participants. As I am the principal investigator on this grant, all funding comes initially to my institution, which must set up agreements with the other sites to distribute to them the funds they are owed.

The e-science pilot projects, some of which fall under different research councils' remits, all interact with the UK e-science core programme. The result has been a stimulating but demanding schedule of meetings and closely coordinated research activities. Unfortunately, the UK academic system has no clear lines of command on such projects, which means that we end up muddling through critical decisions based on a large dose of fudge and what we hope to be academic "goodwill".

Although quite a few of my colleagues still exhibit this old-fashioned virtue, it is egregiously absent in the positions adopted by a growing number of universities. I have just moved from Queen Mary, University of London, to University College London in what has become a protracted and messy affair, with Queen Mary's management attempting to retain assets (kit and funds) that were already committed to the success of a number of publicly funded projects, such as RealityGrid.

I have had to spend a large amount of time since June 2002 handling matters associated with this transfer. Many of the problems result from the funding agencies' lack of well-defined and coordinated policies relating to the movement of academics. If the transfer market for academics continues in its present, almost soccer-like, form, it may become essential for academics to hire agents who can handle the management of their business affairs, which typically include personnel, grants, equipment and intellectual property agreements.

Since arriving at UCL, I have been dogged by a burgeoning set of management and finance-related problems. It took Queen Mary about five months to prepare final expenditure statements on transferring EPSRC grants and for the relevant grants to be reannounced at UCL. And a "global" intellectual property deal for the project was incomplete before the legal team at Queen Mary ceased work on it in June 2002. More than a year later, it is getting significant attention again, at UCL, but it has been slowed by being linked to the budgets to be paid to partner institutions.

The time spent on recovering control of these EPSRC and other grants is, of course, additional to the major demands made by scientific project management. Indeed, with all these administrative headaches to attend to, one might ask if a project such as RealityGrid is worth the hassle. I believe it is. As the main UK e-science project dedicated to high-performance grid computing research, it has pioneered the development of computational steering so that a scientist can interact with a supercomputer-based simulation in real time. One of the projects we are working on is the linking, through grid technology, of the two UK national supercomputing facilities - at Daresbury (HPCx) and Manchester (CSAR) - that are embedded in the UK e-Science Grid with the Extended Terascale Facilities in the US - Jin essence, the "TeraGrid".

This is exciting big science, although simply managing the undertaking is a Herculean task. But that excitement sometimes seems in danger of being lost in the administrative quagmire that has enveloped me over the past two years.

Peter Coveney is director of the Centre for Computational Science in the department of chemistry at University College London. For more information, visit: www.realitygrid.org