Size Estimation for C Programs with Different Applications
Cost estimation is an important aspect of the management of software development projects. Most cost estimation models require an estimate of the size of the program to be developed. Accurate size estimation is thus critically important. Earlier work has shown that weighted counts of conceptually unique variables (where each variable is weighted by the number of functions in which it is referenced) are good predictors of the final size of Pascal and C programs. Simple equations using only that information give acceptable accuracy, even across a range of sizes and applications. For Pascal programs, the accuracy is acceptable by the end of high level design; for C programs, it is acceptable by the end of detailed design. In this study. C programs are classified by application, instead of being treated as a single class. Different estimation equations are then used for programs in each class. Overall estimation accuracy improves at each stage of design. A broad classification scheme is sufficient to give estimates of acceptable accuracy by the completion of high level design.
keyword CodesD.2.8 K.6.1.
KeywordsSoftware Engineering, Metrics Project and People Management
- 1.Albrecht A., 'Measuring application development productivity', Proc. IBM.Applications Development Symposium, October 1979, pp 83 - 92Google Scholar
- 2.DeMarco T., 'Controlling Software Projects', (Yourdon Press, 1982 )Google Scholar
- Dunsmore H., Wang A., 'A step toward early size estimation for use in productivity models', Proc. 1985 National Joint Computer Conference on Software Quality. Williamsburg VA, March 1985Google Scholar
- 4.Campbell R., Conte S.D., Rathi M.K.. `Early prediction of software size and effort', Purdue University Technical Report SER.C-TR-10-P. March 1988Google Scholar
- 5.Rathi M.K.. 'An Objective Methodology for Early Software Size Estimation'. PhD thesis. Purdue University, August 1988Google Scholar
- 8.Lokan G.J., Conte S.D., 'Size estimation for Pascal programs', Purdue University Technical Report SERC-TR-129-P, December 1992. To appear in Journal of Computer and Software Engineering Google Scholar
- 9.Lokan C.J., 'Early size prediction for C programs'. Purdue University Technical Report SERC-TR-139-P,.lune 1993Google Scholar
- 10.Conte S.D., Dunsmore H. Shen V., `Software Engineering Metrics and Models', (Benja.minCummings, 1986 )Google Scholar