Making Resource Decisions for Software Projects

  • Authors:

  • Norman Fenton;William Marsh;Martin Neil;Patrick Cates;Simon Forey;Manesh Tailor

  • Affiliations:

  • University of London and Agena Ltd;University of London and Agena Ltd;University of London and Agena Ltd;University of London and Agena Ltd;University of London and Agena Ltd;University of London and Agena Ltd

  • Venue:
  • Proceedings of the 26th International Conference on Software Engineering
  • Year:
  • 2004

Quantified Score

Hi-index 0.00

Visualization

Abstract

Software metrics should support managerial decisionmaking in software projects. We explain how traditionalmetrics approaches, such as regression-based models forcost estimation fall short of this goal. Instead, wedescribe a causal model (using a Bayesian network)which incorporates empirical data, but allows it to beinterpreted and supplemented using expert judgement.We show how this causal model is used in a practicaldecision-support tool, allowing a project manager totrade-off the resources used against the outputs(delivered functionality, quality achieved) in a softwareproject. The model and toolset have evolved in a numberof collaborative projects and hence capture significantcommercial input. Extensive validation trials are takingplace among partners on the EC funded projectMODIST (this includes Philips, Israel Aircraft Industriesand QinetiQ) and the feedback so far has been verygood. The estimates are sensible and the causalmodelling approach enables decision-makers to reasonin a way that is not possible with other projectmanagement and resource estimation tools. To ensurewide dissemination and validation a version of thetoolset with the full underlying model is being madeavailable for free to researchers.