An object of this meeting is to define and emphasize the best practices that allow stellar photometry done with array detectors to be as accurate as possible. We dream of applying simple and clearly justified principles; in reality, we find the equipment has physical and technical properties that interfere with each other. Photometry is a method, a metrology; so we first have to define the desired accuracy. Some will be happy with 0.1 magnitude; others need 0.01, or even 0.001 magnitude. The necessary precautions must be matched to these ambitions.