The increasing complexity of computer programs and of the development of systems creates an ever growing need of objective methods for measuring the quality of software. A theory providing this measure to some extent has been developed in the last decade by M.H. Halstead. The only parameters in the developed formulae are the number of operators and the number of operands in the routines. Numerous statistical studies have shown a relatively good correlation between theory and experimental results. The theory is applied in this paper to a set of APL defined functions and the results are compared with those obtained by Halstead for other languages.