| Abstrakt: |
An assessment of the software engineering effort /schedule/expected no of defects etc., are required by Project Managers, Customers, Developers, Systems analysts amongst other stake holders. Poor Estimation during the early stages of the software development life cycle leads to schedule, cost overruns. There are several categories of estimation models ranging from Work Break Down Structure, Parametric Estimation and size-based estimation models which exist for predicting the software development effort. Estimates from many of these models do not converge to actual values. Moreover, many of these models provide an estimate of only the coding phase. Due to volatility of requirements and also due to varying problem and solution complexities, varying productivity ratios, varying size and complexity of code, varying testing scenarios the actual value of effort required to complete the requirements phase or effort required to complete the design phase vary considerably depending upon different conditions encountered during the execution of the project. Also, during project execution, one may encounter defects and would spend time removing these defects. Surplus effort may be encountered while removing the defect during the test phase instead of an inspection phase conducted during the Requirements development phase. Simulation models of expected software development effort have been developed in this paper. It is shown that the expected effort varies considerably and cannot be estimated using a single point estimate or even a estimate within a range. [ABSTRACT FROM AUTHOR] |
