An empirical investigation of the tacit plan knowledge in programming

Reviewer: Joseph L. Podolsky

Programming managers quickly learn that all programmers are not equally productive. Excellent programmers are usually orders of magnitude better at creating code and, further, the code created by these experts is “better code,” which is less costly to modify and maintain. It is important, therefore, to discover what distinguishes experts from novices so that we can either select people with those traits, or train people correctly so that all progammers more closely approximate excellence. This discovery is the long term goal of Yale researchers Ehrlich and Soloway. This article describes one small experimental step toward that discovery. It is an important step, but one with limited use by itself. Thus, the article is valuable reading for students of programming behavior, but of little use to practitioners. The authors postulate that a major difference between expert programmers and novices is that experts have tacit knowledge; i.e., knowledge that they don't even know they have. The authors tested this assumption by giving groups of novices, intermediate programmers, and experts (all in an academic environment) several small PASCAL programs in which lines of code were omitted. The programmers were asked to fill in the blank lines. Some of the text programs were staightforward, testing fundamental progam understanding; some, however, purposely created some ambiguities which could be solved in several ways depending upon the inferences drawn from the code. The experiemtns showed that more experienced programmers do, in fact, use tacit knowledge to make the inferences and, therefore, do a better job of completing the test programs. The next goal of this research is one in which there could be major practical results. The authors hope to somehow specify and catalogue the tacit knowledge which experienced programmers use. If this knowledge can be offered in a form such that it is reusable by everyone, we as an industry can learn from each other and build on each others' experiences, instead of spending our energies “reinventing wheels.”