Although the research needs discussed in Chapters 2 through 4 are appealing and potentially valuable, the difficulty of implementing them is not discussed in this report. Some of them may in fact be unattainable in the manner described here or by methods currently available. Some of them have been suggested previously (see the bibliography) but have not yet yielded to research efforts. It is essential to monitor progress in each of these areas and to be ready to halt unproductive approaches and begin anew if necessary.
At the same time, it is important to identify those research problems that are critical, that is, those that may be so integral and pervasive that failure to make progress in them would endanger an entire research effort. In such critical areas, we should be prepared to seek nearer-term results to at least approach, if not necessarily solve, the problems. Thus we can ensure that some results will actually be achieved and can guard against the mindset that the solution is always just a few years away, if only more time and money are spent. A series of increasingly successful near-term solutions would keep us ahead in a pragmatic, if perhaps not elegant, way. Because the ultimate concern is practical results and not knowledge for its own sake, this economizing of effort and continual focusing on useful results must be made part of the administration of the research effort.
An effective research strategy calls for the incorporation of metrics and methods for validation of implicit or explicit hypotheses. Why is a proposed new technology better than conventional technologies? How much better is it? Effective analysis and convincing demonstration of benefits are important for achieving timely dissemination of process improvements and better technology transfer, as well as in the assessment of appropriate next steps.
Finally, research funding levels should take into account the costs of necessary research infrastructure. A $100,000 grant, for example, may cover the mathematical elements of a software system for manufacturing, but it will not cover the physical elements necessary to test a system. In manufacturing, the algorithms and the immediate technical implementations are only half of the picture.
Common products such as automobiles can have thousands of parts, and modern aircraft and integrated circuits include millions of parts or active elements. Each of these examples takes years to design, requiring the design effort of hundreds or thousands of people located in diverse areas around the world. Complex new products based on information content and their accompanying information-dominated design and manufacturing methods already require us to deal with entirely new scales of complexity. Some products require such levels of precision, delicacy, or cleanliness that people can no longer make or assemble the parts; in some cases, they cannot even see them.