CULTURE AND RESEARCH INFRASTRUCTURE 361
problems include tasks of designing specific proteins in less than a day that can catalyze specific
chemicals or inactivate an influenza virus.
As a vehicle for pursuing its mission, DARPA typically uses Broad Agency Announcements. Some
are focused on achieving a specific technical capability (e.g., a program that will develop technology for
synthesizing, within 24 hours, an arbitrary 10,000-oligonucleotide sequence in quantity), whereas oth-
ers are more broadly cast. In many cases, these programs target private industry as well as the more
engineering-oriented academic institutions.
10.3 BARRIERS
Because work at the BioComp interface draws on different disciplines, there are barriers to effective
cooperation between practitioners from each field. (In some cases, “each field” is more properly cast as
the contrast between practitioners of systems biology and practitioners of empirical or experimental
biology.) This section describes some of these barriers.^63
10.3.1 Differences in Intellectual Style,
It is almost axiomatic that substantial progress in any area of intellectual inquiry depends on the
excellence of work undertaken in that area. On the other hand, differences in intellectual style will affect
what is regarded as excellence, and computer scientists and biologists often have very different intellec-
tual styles.
The existence of shared intellectual styles tends to increase the mutual understanding of colleagues
working within their home disciplines, a fact that leads to more efficient communication and to shared
epistemological understanding and commitments. However, when working across disciplines, lack of a
shared intellectual style increases the difficulties for both parties in making meaningful progress.
What are some of the differences involved? While it is risky (indeed, foolhardy) to assert hard and
fast differences between the disciplines, an examination of the intellectual traditions and histories
associated with each discipline suggests that practitioners in each are generally socialized and edu-
cated with different styles.^64 Over time, these differences may moderate as biology becomes a more
quantitative discipline (indeed, a premise of this report is that such evolution is to be encouraged and
facilitated).
10.3.1.1 Historical Origins and Intellectual Traditions
Many differences in intellectual style between the two fields originate in their histories.^65 Computer
science results from a marriage between mathematics and electrical engineering—although it has
evolved far from these beginnings. The mathematical thread of computer science is based on formal
problem statements, formulating hypotheses (conjectures) based on those statements, and generating
formally correct proofs of those hypotheses. Most importantly, a single counterexample to a conjecture
invalidates the conjecture. Note also that formal proofs often entail problems that are far from reality,
because many real problems are simply too complex to be represented as formal problem statements
that are at all comprehensible. Research in mathematics (specifically, applied mathematics) often con-
(^63) An early perspective on some of these barriers can be found in K.A. Frenkel, “The Human Genome Project and Informatics:
A Monumental Scientific Adventure,” Communications of the ACM 34:40-51, 1991.
(^64) An interesting ethnographic account of life in an academic biology laboratory is provided in J. Owen-Smith, “Managing
Laboratory Work Through Skepticism: Processes of Evaluation and Control,” American Sociological Review 66(3):427-452, 2001.
(^65) Some of this discussion is inspired by G. Wiederhold, “Science in Two Domains,” Stanford University, March 2002, updated
February 2003. Unpublished manuscript.