Let me call your attention to an excellent summary of a very useful and creative workshop on the ‘History and Future of Infrastructure: Lessons for New Scientific Cyberinfrastructures’ held in Ann Arbor in September 2006. The summary report on the workshop is authored by Edwards, P. N., Jackson, S. J., Bowker, G. C., and Knobel, C. P., and entitled ‘Understanding Infrastructure: Dynamics, Tensions, and Design’ (January 2007) and is available online at: http://www.si.umich.edu/InfrastructureWorkshop/

This is an important overview for anyone seriously interested in the study of e-science, e-research and related topics. This conceptualization has already shaped the naming of NSF’s programme on Cyberinfrastructure, as well as discussion in the UK about e-Infrastructures.

However, as a participant in the workshop and a reader of the summary, I found the report to be most useful in stimulating questions about the concept of ‘infrastructure’ as applied to e-science and e-research. I can see immediate value in the concept: It moves away from the specific technologies of e-Science, such as the Grid, to the entire ensemble. It also enables comparison with other ‘infrastructures’, from canals to telecommunication systems. Some of my reservations are:

1. Why is it more useful to view the electronic media supporting scientific research as ‘cyberinfrastructure’ as opposed to technology, or information and communication technologies, more specifically?

2. Does the concept of infrastructure imply a certain stage of development, which might not apply to the area of e-research, given it emerging nature. For example, I think it is useful to understand the transition of the Internet from a ‘new technology’ to a taken for granted ‘infrastructure’ in some regions of the world. If this is the case, do we need a theory of routinization as developed within work on innovation v a theory of infrastructure?

3. Are we accepting a policy term, used by the NSF, as a scientific concept? Probably. What might be useful for policy and practice might not be useful for analysis.

4. Does infrastructure imply more coherence and systematization that warranted? Maybe. In the 1970s, my colleagues and I resisted the use of the term ‘computer systems’ since we often found computers in organizations to be organized in much less systematic ways than implied by such a term. Computers were sometimes in the closet. Sometimes they were dysfunctional in other ways. Instead we spoke of the ‘computer package’ of people, techniques and equipment surrounding the provision of computing services.

I don’t wish to detract from the report, which is excellent, but push a possibly useful debate and more critical perspective on the very concept of infrastructure. I’d welcome responses.