I’m not an engineer but I’m guessing that a perennially fascinating and relevant debate is how to test a risky new product. In working through the 616-page blur of technical matter in Fluid Fuel Reactors, I came across a pithy, cogent discussion of that issue in relation to a particularly exotic stream of reactor design. The words below get into too much detail for me to use in the book, but I’ll move forward alerted to the distinctions covered:
Although many have argued that the shortest route to economic power will be achieved by eliminating the intermediate-scale plants, most experts believe that eliminating these plants would be more costly in the long run. To quote from a speech by Dr. A. M. Weinberg, while discussing large-scale reactor projects: “The reactor experiment—a relatively small-scale reactor embodying some, but not all, of the essential features of a full-scale reactor—has become an accepted developmental device for reactor technology.” An alternative to the actual construction of experimental nuclear reactors has been proposed which consists of the development of reactor systems and components in nonnuclear engineering test facilities, zero-power critical experiments, and the testing of fuel elements and coolants in in-pile loops. This approach, although used successfully in the development of various solid-fuel coolant systems, is not completely applicable to circulating-fuel reactors because of the difficulty of simulating actual reactor operating conditions in such experiments. … A second aspect of circulating-fuel reactors, which precludes relying solely on engineering tests and in-pile loops, is the close interrelation of the nuclear behavior and the operational characteristics of the fuel circulation system, which can be determined only through construction and operation of a reactor.
Lane, James A., H. G. MacPherson & Frank Maslan. 1958. Fluid Fuel Reactos. Addison-Wesley, Reading, Massachusetts, p. 221.