How can observation of a legacy system be used for design? To answer this question, we observed a teacher doing live coding with a conventional code editor and used the observations to design an editor better suited to this style of teaching. In particular, we found strong evidence that a projectional editor would better meet this need. Reflecting on this experience, we describe two types of requirements which can be inferred from observing a user using a legacy system: hidden requirements, in which users use existing features in unexpected ways, and novel requirements inferred from pain points observed in current system use. / Thesis / Master of Applied Science (MASc)