Planning agile software projects with reduced guess estimation

Abstract

Agile Project Management (APM) is the method of a series of commonly used project management approaches for better handling of uncertainty and unpredictability. In large portion of software projects, accurate planning of project lifetime with Agile is di cult. Since continuously change in requirements occurs as well their incompleteness at project initialization, project plans must be under control and continuously revised. The purpose of this thesis is to help tackling di culties of managing Agile software projects by proposing an improved software planning technique. In Agile projects, planning is mainly based on just guess estimation of global e ort. In the proposed approach, instead of just guess estimation, there is measurement of the size of the product. This measurement is implemented over user stories with COSMIC measurement and veri cation methods. Thus, needed e orts for the projects are calculated. The approach is applied on a lm renting project in order to present its applicability. The results show that, by using the proposed approach, sizes of all pieces of the product can be obtained and by using the measurement results, project plans can be prepared with higher reliability and maintainability. Project plans may be more reliable, because they depend on solid measurement results and they may be more maintainable, because if any of the requirements changes, just that requirement needs to be measured again. Also it is shown that, documentation quality has an impact on the measurement results. As documentation quality changes, measurement results change by depending on the information in the documentation. It is also presented that, identi cation of the quality of the documentation helps to criticize the measurement.