| Author(s) | Aftab Ahmad |
| Affiliation | Department of Computer Science, College of Computer and Information Sciences, King Saud University, PO Pox 5II78, Riyadh II543, Saudi Arabia |
| Title | Software Science Analysis of PROLOG |
| Source | Journal of King Saud University. Computer & Information Sciences. Volume 6, No 1. (1994/1414) |
| Abstract | The theory of software science attempts formularization of software characteristics. It was developed by the late M. H. Halstead of Purdue University during the early 1970's. During the past years, large numbers of empirical and psychological studies were performed to validate and investigate capabilities of proposed software metrics. Most of these studies supplied evidence to support the theory. The on-going process of verification and refinement of theory initiated its application in wide areas includ-ing: estimating co plexity of software maintenance and relating software errors to the theory. Within the fame of these investigations, a great interest has been shown towards application of software science t programming languages for evaluation of their relative levels and testing of length equ-ation. In this reg d an extensive published work exists covering experimentation on various program-ming languages. owever, all previous studies have been mainly concerned with application of software science to convent onal 'Procedural Languages'. It obviously triggers motivations to explore applicability of software scienc to newly emerging 'Declarative Languages'. The work in t is paper reports on a pioneer investigation into the application of software science to a fifth generation eclarative language like Prolog. Various tools and databases have been developed which provide on- ine software metrics research environments and data-collection for declarative lan-guages including olog. These tools were used to analyze large number of Prolog programs collected from published lite ature and university environments. Two fundamental counting rules and their possi-ble interaction wer studied for Prolog programs. The results from experimentation clearly indicate to the significant fact that software science can be successfully applied to Prolog. These results have nicely ver-ified the length equ tion. The study concludes that Prolog has higher language level than all other prog-ramming languages reviously studied. This prediction strongly confirms the common intuition on Prolog level relative to the ther procedural languages. |
