Calibrated Estimation Model for a Maintenance Project

N.K. Bhandari and H. Goswami (India)


Function Points, Maintenance, Estimation, and Calibration.


Effort and size estimation for legacy products under maintenance phase is very challenging as most of legacy products were developed several years ago when software estimation techniques were not mature; these projects were often estimated using the rule of thumb. In this paper, we propose a simple yet effective model to estimate size and effort for maintenance of a legacy product based on Function Points, which are calculated from the Lines of Code (LOC) using a reverse engineering technique. Subsequently these Function Points are calibrated in order to accommodate factors such as product/domain knowledge and learning curve characteristics. The proposed model builds on the fact that the maintenance effort for a domain knowledge intensive project is substantially different than the maintenance effort for normal projects. It also takes into account the fact that descend on learning curve is steeper in the case of domain intensive projects. We validate outcome of the analytical model with measurements from a real-world maintenance project of a legacy product that heavily depends on domain knowledge and learning curve characteristics. The validation shows that the results from this analytical model and the real world data are in close synergy, which emphasizes the effectiveness of the model.

Important Links:

Go Back