selected. Those concerns have eroded confidence among policy makers and the research community that large research facility projects are being ranked on the basis of their potential returns to science, technology, and society.
To address the concerns regarding NSF’s process for identifying, approving, constructing, and managing large research facility projects, the committee makes the following recommendations:
1. The National Science Board should oversee a process whereby the National Science Foundation produces a roadmap for large research facility projects that it is considering for construction over the next 10-20 years.
Broad inputs from the scientific community must form the basis for the roadmap.
The roadmap should take into consideration the need for continued funding of existing projects and should provide a set of well-defined potential new project starts for the near term (0-10 years). These projects should be ranked against other projects expected to be funded in a given year and according to where they are positioned in time on the roadmap. Projects further out in time (10-20 years) will necessarily be less well defined and ranked qualitatively to yield a vision of the future rather than a precise funding agenda, as is the case for the earlier years.
Different categories of overlapping criteria, described briefly in the bullets below, need to be used as one moves from comparing projects within a field to comparing projects in a directorate or in the entire NSF. At each level, the criteria used in the previous level must continue to be considered.
Within a field (as defined by NSF division) or interdisciplinary area: scientific and technical criteria, such as scientific breakthrough potential and technological readiness.
Across a set of related fields: agency strategic criteria, such as balance across fields and opportunities to serve researchers in several disciplines.
Across all fields: national criteria that assess relative need—such as which projects maintain US leadership in key scientific and engineering fields or enable the greatest numbers of researchers, educators, and students.
See box on page 4 for a more in-depth discussion of the proposed criteria.
A key constraint that must be imposed in the final stages of development is that the roadmap must reflect a reasonable projection of the