Background:

Estimating (predicting, forcasting) of effort required for software development work is difficult and quite inaccurate. The typical effort estimate is about 30% too low. In spite of more than 40 years on work on this developing support (models) for effort estimation, there has not been much improvement and effort estimates are still inaccurate and over-optimistic. This master thesis is based on the idea that we may have to live with inaccurate effort estimates, but should be better to assess how inaccurate (uncertain, risky) they are. For that purpose, researchers at Simula have developed an approach that uses historical estimation accuracy to predict the uncertainty of the current effort estimate.

What you will do:

The work will consist in implementing and evaluating the approach (as outlined in for example Jørgensen and Sjøberg 2003)) proposed by the researchers at Simula as a prototype of a web-based tool, to be used by software developers at no cost. Currently, we have a lot of software professionals visiting our web-page and we expect them to use the tool and be willing to share their experience. Alternatively, we may invite a few companies and ask them to try the tool. The student should be interested in programming, experience with web technologies such as Ruby on rails and Groovy is an advantage.

What you will learn:

• Web-programming (the tool has to be robust, maintainable and user friendly)
• Methods for empirical evaluation of software technology
  
• Methods for effort estimation
  

References:
Jørgensen, M. and Sjøberg, D. (2003). "An effort prediction interval approach based on the empirical distribution of previous estimation accuracy." Journal of Information and Software Technology 45(3): 123-136.