Web development effort estimation using analogy

Although estimating the effort required in developing Web applications is a difficult task, accurate estimates of development effort have an important role to play in the successful management of Web development projects. In software development work to date, emphasis has focused on algorithmic cost models such as COCOMO and function points. Two disadvantages of these models are firstly, the need for calibration of a model for each individual measurement environment and, secondly, the variable accuracy levels achieved even after calibration. The paper describes the use of estimation by analogy to calculate the development effort of Web applications. Two datasets containing empirical Web development data were used in the case study. One set contained data relating to forty-one novice developers, the other to twenty-nine experienced developers. The ANGEL tool supporting the automatic collection, storage and identification of the most analogous projects was used as a basis for estimating effort required for a new project. Results show estimation by analogy to be a promising alternative to algorithmic techniques.