ÿÈÕ³Ô¹Ï

In an undertaking such as the U.S. Cyberinfrastructure Initiative, or the UK e-science programme, which spans many years and comprises a great many projects funded by multiple agencies, it can be very difficult to keep tabs on what everyone is doing. But, it is not impossible. In this paper, we propose the construction of ontologies as a means of monitoring a research programme's portfolio of projects. In particular, we introduce the "virtual laboratory ontology" (VLO) and show how its application to e-Science yields a mapping of the distribution of projects in several dimensions of the "collaboration space." We sketch out a method to induce a project mapping from project descriptions and present the resulting map for the case at hand. What the map suggests is that the UK's e-Science programme so far has remained very "data-centric". Apart from methodological bias, two hypotheses could account for this focus: distributed databases are of central importance to a wide array of science and engineering fields, and tools for federation, annotation and facilitated access form a logical priority in middleware development (H1); there is a preference for organizing projects that involve dyadic interactions with a research facility that requires no intervening human agency, and an aversion to undertaking contracts for collaborative work among research groups that would transcend institutional or organizational boundaries (H2). Further studies that would make use of ancillary information, including interviews with principals in the formation of the UK's e-Science core program, will be needed to throw light on the validity of these or still other potential explanations. Be that as it may, this paper shows that the proposed mapping approach to be informative as well as feasible, and we expect that its further development can prove to be substantively useful for future work in cyber-infrastructure-building.