A case study of run‐time errors in Pascal programs

The results of a case study in which over 100,000 Pascal program executions were monitored for run‐time errors are reported. A large number of run‐time errors in a wide variety of categories were observed. The data reported provided insight into the use and misuse of the features of Pascal by a large population of programmers. Some implications of these statistics on compiler implementation and programming language design are discussed. The number and variety of errors detected suggests that run‐time checking mechanisms are more important and useful than is generally recognized, judging by the incompleteness of such mechanisms in many compilers.