Tuesday, August 31, 2010

DARPA and NSF

The one topic that others have focused on most in conversations with me has been a comparison between DARPA and NSF funding of science. As scientists know, NSF is peer-review oriented and programs tend to cover a traditional area of science and have a relatively long life-span (except for term-limited special programs).

The, what some would say is, "traditional DARPA" is quite different. DARPA programs are approved by the DARPA Director so they take on the character of that person and their interests. Nevertheless, there has been a typical model whereby programs are "sold" to the Director using a presentation that follows the Heilmeier Catechism (see result of Google search on this term). Because NSF does a good job of covering most of science for the US, DARPA can afford to pursue more risky, but higher payoff ideas. The Internet is the best example. No need to recount that history, but you get the idea. Others would like to create programs with that kind of impact, hoping it wasn't just a fluke. Therefore, they want to know how DARPA works (or worked, in that case).

When I hear what other countries are interested in creating, I get the strong feeling that they are less interested in the general health and progress of science than in quick-win programs that could have high payoff. Therefore the interest in hearing about how DARPA programs are created and run. My main contribution to that discussion has been to emphasize that there are two really important factors besides meeting the goals of the Heilmeier Catechism: a program manager with a strong vision and a DARPA Director who buys into that vision without re-directing or mis-directing it.

DARPA used to be known as an agency of 100 program managers united by a common travel office. The point this tries to capture is that DARPA PM's seem to work best when they survey in person the research capabilities in their interest area and create a vision from what they learn. You can't create a program without knowing who would be able to do the required research, and you also need feedback on your own ideas to help clarify and focus them, specifically to answer the catechism questions.

Probably the hardest lesson for other countries in learning about this process from my point of view has been to appreciate the type of person who would be a good program manager in a DARPA-like program and to actually find people with the skills needed. Such people are rare. In my own experience, I have been surprised to find that my cultural anthropology training has been more important than my computer science training because it allowed me to understand the state of art in an area culturally. That is, who the main players are, what the social structure is (conferences, universities, research labs, etc.) and what the main problems of the field are that keep people up at night. Answering these questions requires being truly objective so as not to bias your observations with your own opinions or desires. This kind of objectivity is what ethnologists in the field of cultural anthropology really try to achieve, although not all are successful even there.

I would be very interested in reading comments by anyone else co-trained in both cultural anthropology and a science.

Initial

After retiring from a career in science funding and management (see my LinkedIn posting under "Gary W. Strong"), I decided to start a blog because of a felt need to share my experiences, warts and all.

While the US government science funding program is remarkable in its successes, anything can be improved. Comments from anyone on the subject are more than welcome. We are looking to create a conversation rather than simply do a brain dump.

Having taught a couple of workshops on research management in Saudi Arabia has also clarified to me what some of the lessons were that I learned but took for granted. When you explain to someone else how you did your job, you quickly find out that you actually have strong opinions based on experience that could be valuable to others.

While my very best job was at the National Science Foundation, I also worked at DARPA as a PM and at the Department of Homeland Security in charge of Behavioral research program plans and budgeting. Most positions also allowed me to interact with nearly all of the other Federal agencies in one way or another in matters of research funding cooperation.

This blog won't be a life story. I will post items as they occur to me in probably a more-or-less random order. If you just found me here, thanks, and please let me know what you think.