In lieu of an abstract, here is a brief excerpt of the content:

  • Reviews of Recent Publications by Grier and Hemmendinger
  • David Hemmendinger (bio)
Thomas J. Misa and Jeffrey R. Yost, FastLane: Managing Science in the Internet World, Baltimore: Johns Hopkins University Press, 2015. 224 pp. ISBN: 9781421418681.

The US National Science Foundation (NSF) FastLane system handles all its research grant proposals, reviews, budgets and project reports. Its development began in 1994, its use became mandatory in 2000, and its current user-help document reports its use by 400,000 people annually.

Tom Misa and Jeff Yost, both of the Charles Babbage Institute, studied the development and use of FastLane in a project that ran from 2007 to 2011, supported by NSF research grants. They interviewed over 800 participants, mostly researchers; about 10 percent were NSF staff members. Half of the interviews were done in person, and half online. About 80 percent of the interviews are publicly accessible at the Babbage Institute—the authors suggest that their dataset may be the largest available on the development and use of a complex computing system.

The authors study FastLane from several perspectives: as a government information technology project, as e-commerce that developed in parallel with the World Wide Web and the Mosaic graphical web browser, and as a software engineering project. They look at how it changed both researchers and the NSF.

The book describes the early stages of automation at the NSF. Connie McLindon came to the NSF in 1980 from the Defense Advanced Research Projects Agency (DARPA) and, in the early 1980s, upgraded NSF servers and introduced network email, which only the computer science program officers previously had. Erich Bloch, who became director in 1984, promoted automation, and McLindon had an important role in developing EXPRES, a system for email submission of multimedia research proposals. This system did not impose a document standard, but had translation programs to convert a variety of formats to a single intermediate format. It was ambitious, and also premature at a time when there were multiple email and networking standards. Nonetheless, it helped prepare for FastLane.

The experimental FastLane project that began in 1994 coincided with the emergence of the World Wide Web and, in particular, with the Mosaic graphical browser developed at the Illinois National Center for Superconducting Application, with heavy NSF support. It shared important capabilities with e-commerce programs such as those used by Amazon: encryption for security, a database, and dynamic webpages that loaded information from the database. It used the Perl scripting language with PostScript forms and file uploads, something initially possible only with [End Page 82] the Netscape browser that succeeded Mosaic. It also required researchers to upload their proposals as Adobe Portable Document Format (PDF) files, starting in 1996, when PDF document production was not widely or cheaply available. It required Adobe products or free alternatives that might not produce PDFs acceptable to FastLane. Despite these difficulties, FastLane enabled researchers to avoid last-minute pressure to package two dozen copies of proposals and rush them via priority mail or even hand-carry them to the NSF. Users did find that FastLane got swamped by last-minute submissions, as some practices never change.

Despite frustrations due to PDF difficulties and an overloaded system, PIs (principal investigators) generally liked FastLane, though crashes of an overloaded system led some to think that requiring its use in 2000 was premature. The overall satisfaction with FastLane may have been a result of what the book calls “value-laden design,” in which key NSF values were designed into the system. These included open access by small colleges as well as research universities, security, reliability, and protection of merit reviews. The designers also got frequent user feedback, both from formal committees and from a network of researchers and university administrators. They sought to make FastLane accessible to both occasional users and university grants officers who dealt with it daily—one goal was not to change the interface frequently, so that the former group would not have to learn a new system on each use. Nonetheless, the authors point out that PIs at small institutions, such as HBCUs (Historically Black Colleges and Universities), and underfunded institutions in the EPSCoR project (Experimental Program to Stimulate Competitive Research) were likely...

pdf

Share