A Study of Applicability of Complexity Measures Export

@article{Davis88a, abstract = {A study of predictive value of a variety of syntax-based program complexity measures is described. Experimentation with variants of new chunk-oriented measured showed that one should judiciosly select measurable software attributes as proper indicators of what one wishes to predict, rather than hoping for a single, all pirpose complexity measure. This study has shown that it is possible for particular complexity measures or other factos to serve as good predictors of some properties of program but not for others. For example, a good predictor of construction time will not necessarily correlate well with the number of error occurences. Halstead's effort measure (E) was found to be a better predictor than the other two nonchunk measures we evaluated: McCabe's V(G) and lines of code, but at least one chunk measure predicted better than E in every case.}, author = {Davis, John S. and Leblanc, Richard J.}, citeulike-article-id = {3013381}, journal = {IEEE Transactions on Software Engineering}, keywords = {complexity, evolution, science, software}, month = {September}, number = {9}, pages = {1366--1372}, posted-at = {2008-07-17 05:23:24}, priority = {2}, title = {A Study of Applicability of Complexity Measures}, volume = {14}, year = {1988} }

TY - JOUR ID - Davis88a L3 - citeulike-article-id:3013381 N2 - A study of predictive value of a variety of syntax-based program complexity measures is described. Experimentation with variants of new chunk-oriented measured showed that one should judiciosly select measurable software attributes as proper indicators of what one wishes to predict, rather than hoping for a single, all pirpose complexity measure. This study has shown that it is possible for particular complexity measures or other factos to serve as good predictors of some properties of program but not for others. For example, a good predictor of construction time will not necessarily correlate well with the number of error occurences. Halstead's effort measure (E) was found to be a better predictor than the other two nonchunk measures we evaluated: McCabe's V(G) and lines of code, but at least one chunk measure predicted better than E in every case. IS - 9 JF - IEEE Transactions on Software Engineering EP - 1372 TI - A Study of Applicability of Complexity Measures VL - 14 SP - 1366 KW - complexity KW - evolution KW - science KW - software AU - Davis, John AU - Leblanc, Richard PY - 1988/September// ER -