Our activities in software engineering typically fall into one of three categories, (1) to invent new phenomena, (2) to understand existing phenomena, and (3) to facilitate inspirational education. This paper explores the place of empirical software engineering in the first two of these activities. In this exploration evidence is drawn from the empirical literature in the areas of software inspections and software cost modelling and estimation. This research is then compared with the literature published in the Journal of Empirical Software Engineering. This evidence throws light on aspects of theory derivation, experimental methods and analysis, and also the challenges that we face as empirical software engineering evolves into the future.