Issues important for new ultrashort-pulse-measurement techniques include the generation of theoretical example traces for common pulses, validity ranges, ambiguities, coherent artifacts, device calibration sensitivity, iterative retrieval convergence, and feedback regarding measurement accuracy. Unfortunately, in the past, such issues have gone unconsidered, yielding long histories of unsatisfactory measurements. We review these issues here in the hope that future proposers of new techniques will consider them without delay, and, as an example, we address them for a relatively new technique: self-referenced spectral interferometry.