In this talk, two examples are given of the process of translating user requirements into optimization and assessment tools. In the first place, recent work on the effects of aberrations on the perceived image quality of visual instruments is reviewed. This allows the assessment of a visual system in terms of expected loss of contrast and resolution as a function of aberration, and also the formulation of an image quality metric suitable for automatic optimization. The second example concerns the extraction of accurate spectroscopic information from pushbroom imaging spectrometers. It is shown how the user requirements for calibration translate into spectral and spatial uniformity of response, and further to the complete absence of spectral and spatial distortion, as well as to the minimization of the variation of the LSF width in both directions, spatial and spectral. Techniques for accomplishing this in practice, both in terms of merit function and in terms of fabrication and assembly, are also discussed.