Usually legacies act as strong amalgamating forces and are thus beneficial to systems and theories. However, as time passes through, even legacies must be reviewed in order to benefit from changes. Although several sound models has been produced by software engineers in the last 25 years, a large amount of software is still produced and maintained using aged paradigms and methods (if any) and COBOL/FORTRAN generation tools. The underlying idea seems to be "Why change it if it works?". Nevertheless it seems too that time for changes has come. The OO paradigm is really strong and useful and has been showed appropriate for almost any conceivable software system. Besides, actual programmers and analysts are formed under this paradigm. So it seems foreseeable a consistent conversion to OO paradigm. What still doesn't seem to be clear is how adequate are the actual OOP languages to support the paradigm, and what is the comparative effort involved in mastering them and their associated environments. This paper reports the heuristics and observations resulting from a simple experiment formulated to explore how adequate are the languages C++, Eiffel, and Ada 95 to implement OO designed systems.
/lp/association-for-computing-machinery/e-pluribus-unum-oopl-selection-PdL5M9DlUE