on Public service and
Members of the computer science community should become more
involved in public service by becoming program managers at federal
agencies, the opportunities and benefits of which are outlined here.
There has BeeN increasing con- cern about the availability of support for basic research in computer science.
2 Al- though there are always
factors such as budget priorities and
federal funds availability, based on
my experience as a program manager
at the U.S. Defense Advanced Projects
Research Agency (DARPA) I believe
there is an additional factor that is
often missed. Neal Lanea has spoken
and written on related issues for science and technology in general; here,
I focus on computer science.
The issue is the strength of the computer science representation “at the
table” when priorities are decided and
resource allocations are determined.
The only way to change the resources
allocated to computer science as a
discipline is to ensure the best new
ideas are represented passionately, effectively, and repeatedly. This means
going beyond the advocacy role of representative organizations such as the
Computing Research Association and
voluntarily becoming public servants:
to become agents for change.
Program management reduces risk, increases reward.
Risk vs. Reward
What can You Do?
Federal agencies such as DARPA and
the U.S. National Science Foundation
(NSF) are constantly looking for new
ideas and people to make them hap-
a See http://www.nsf.gov/news/speeches/lane/
pen. Making ideas happen is the role
of public servants such as DARPA program managers and NSF program directors, and these ideas are often far
beyond what the risk/reward threshold of the private sector can support.
The figure here depicts a conceptual model of the effect that good
program management can have on
research risk and reward. The line labeled “Risk/Reward Curve” is meant
to suggest there is a natural relationship between risk and reward, where
low risk leads to low rewards and high
risk is required to achieve the highest rewards. It is the goal of program
management to create an environment in which the most rewarding research is pursued with the least risk.
The rewards are typically clear in the
program manager’s vision, so it is
risk that must be managed most aggressively.
Risks to manage include problem difficulty, researchers, time, and
money. To manage problem difficulty,
intermediate goals can be targeted, or
multiple competitive approaches can
be supported to reduce risk through
a process akin to “portfolio management.” Researchers must be encouraged to be bold, but at the same time