The requirement that the data be in a computer readable format puts strict limits on what an evaluator can do. Not every fact or physical effect can always be squeezed into the constraints of the format. But the evaluator has to do his or her best to get a good representation of reality. When these constraints are too limiting, we can try to extend the formats.

The fact that these evaluated data files are intended to be used for applications also has consequences. The data must be "complete" in some sense; for example, all energies must be represented, even when the data don't cover them all. Otherwise, particle transport calculations would be impossible. Modern evaluations are done by combining the experimental data with nuclear model code calculations to extend or interpolate the available data. In addition, the fact that these evaluated data files are intended to be used for applications adds factors of quality control, revision control, peer review, and data testing that might not be needed for other modes of publication.