Sir

Plans to impose effort-reporting on scientists, as mentioned in your Editorial 'On the paper trail' and News story 'Researchers criticized for poor time-keeping' (Nature 449, 508 and 512–513; 2007), will be difficult to implement. In practice, it is almost impossible to give an accurate estimate of effort, because scientists are rarely off the job, even when asleep. If they are not actually doing a particular task, they are planning the next, or puzzling over the most recent observation. How should that time be counted? In most research, the edges of a project are only known indistinctly. So in many cases it is difficult to know when one has wandered from one project to another or into an unfunded area.

The definition of '100% effort' causes another problem. This metric tends to be based on a 40-hour week, but not many funded scientists can afford to do so little work — 80 hours (200% effort?) is more common. Also, in these days of tight budgets, how should the project director account for the time spent on writing and revising new grant applications? Under the current funding system, a very large fraction of time is spent on this activity.

Finally, it must be recognized that boot-strapping a new project with the funds of an existing project is built into the system. If there is no venture capital available, new grants have to be built on the backs of existing ones. After all, that's how they got funded in the first place. When was the last time a new proposal was funded in the absence of any preliminary data?

There may be some abuse in very large labs with multiple sources of funding. Filling in the blanks on applications and enforcing appropriate effort reports should be the norm. But please don't reduce the time available for research still further by making researchers account for every moment.