APL compared with other languages according to Halstead's theory

Abstract

Some ten years ago M.H. Halstead and his collaborators at Purdue University set up a theory allowing to measure the characteristics of software and their relations.One of these characteristics is the language level allowing to compare the respective effects of these languages on the productivity of programming. This language level has been evaluated for APL by the author (APL81) and by R.A. Mauri and A.H. Williams (APL82), the latter taking into account the distinction in APL between primitive functions and operators. The results are compared and some reflections are given.

Reliance on correlation data for complexity metric use and validation

“…Significant anomalies have arisen in the consistent classification of tokens, particularly for languages which have emerged since the theory was developed [5], [6]. This has resulted in a situation where researchers have often had to use their own counting schemes, introducing a degree of subjectivity into what is said to be an objective quantification.…”