Information Technology Research

Abstract

In a recent editorial ( Science 's Compass, 6 Aug., p, [833][1]), D. Allan Bromley comments on “appropriate” information technology (IT) research. I agree fully with the importance he attaches to fundamental research and to continued support of basic sciences that would lead to new advances that would have an impact on IT. However, contrary to some of his comments, that is also what the President's Information Technology Advisory Committee (PITAC) report recommended. That committee simply had a different (broader and, I believe, sounder) view of basic IT research than what Bromley presents. Bromley draws far too narrow a boundary around the sciences that are basic to research in IT, identifying them largely with the physical sciences underlying IT hardware. Without denigrating the importance of continuing development of faster and more reliable hardware, I believe that the most critical and fundamental research questions in IT for the coming decades, as PITAC argued correctly, concern the basic principles of the organization of complex information-processing systems, at both hardware and software levels. These are not primarily physical-science problems, but arise wherever high levels of system complexity are found. These organizational problems are, for example, central to understanding the evolution of biological systems and adaptive and learning systems in general, including IT systems. There is a whole history of disaster in IT research that ignored such organizational problems (for example, research on the design of parallel general-purpose systems), and the computer industry lagged for decades in recognizing the importance of relevant basic software research. Theory and design of the organization of both hardware and software are crucial to the success of the enormous, highly decentralized systems that are already becoming an integral part of our whole industrial, governmental, military, and social communication structure and which are today given real-world tasks that are way beyond their capabilities for rapid, accurate, intelligent, and safe performance. Bromley is mistaken in saying that House bill H.R. 2086 is simply supporting competitive industries. Solutions to fundamental software problems of designing ultra-complex systems are extremely important for the future security and competitiveness of the nation, they are not of burning interest to industry today, and they are largely pursued in universities. The PITAC report also recommended a study of the social implications of IT. These equally basic and important research problems must also be addressed vigorously by the public sector. [1]: /lookup/doi/10.1126/science.285.5429.833