Abstract Prognostic Soil-Plant-Atmosphere (S-P-A) system models should be made as robust as possible; that is, parameters should be independent of variables and constant during applications with similar initial conditions (no calibration). However, avoiding calibration requires more cooperation among neighboring disciplines with a commensurate increase in time and effort for development. An example is provided to illustrate the unfortunate consequences of inadequate cooperation among disciplines. Perhaps institutional disciplinary structures are becoming too overbearing, so that scientists are constantly tempted to take shortcuts that make cooperation with colleagues in different disciplines more unlikely. If practitioner scientists would rekindle a close relationship with philosophers of science, who study the actual conduct of science, a more realistic perspective might evolve to reduce institutional structural impediments and connect scientists more closely to the larger community that needs and supports them. Certainly persistence would be required to begin and sustain such a process of change.