This article reports on a study of the knowledge practicing software professionals use in their jobs.The study consisted of a survey asking respondents to rate 75 different kinds of knowledge (generally reflecting the content of typical university computing curricula) on four dimensions: coverage during formal education, current level of knowledge, importance on the job and general personal importance. The author also computed a combined importance rating from job and personal importance. Of most relevance for the role of mathematics in computing, discrete math topics were roughly spread across the middle two quartiles on combined importance -- i.e., they were not the most important, nor the least. Topics from continuous math were solidly in the bottom quartile.
The importance of this paper is that it, or others like it, will inevitably become reference points for any discussion of math's place in computing education. Different people will have different interpretations of the data and different views on whether it is relevant at all, and any study of this sort is easy to criticize. I hope that those who disagree with Lethbridge's results will carry out their own alternative studies. But in the end, no-one who wants to argue for math's (or anything else's) importance in computing education, and who wants to be taken seriously, can afford to ignore this work and its successors.