An empirical evaluation of 4GL and CASE tool use on software engineering productivity models in development of business type systems

Putnam and Boehm suggested techniques for estimating effort, manning levels, time and costs of application software development before fourth generation languages (4GL) and computer aided software engineering (CASE) tools came into fairly wide use. This study used representative software development projects to determine if these models were appropriate for estimating software development resource requirements under changed development methodology. Empirical data was used to evaluate Putnam's model and Boehm's COCOMO models, calibrate those models and determine if the models can be useful in estimating effort and time requirements. Essential conclusions were that the models as presented were useful only if properly calibrated to the development environment and tools used. Software engineering productivity factors, variables and models for prediction in 4GL and CASE environments are suggested. Parameters for the models were calculated and also evaluated for 1982, 1987 and 1994 data sets published by others.