and compare each criteria separately. In this research we will focus on customisability and modifiability, this because Everest indicated that these are the two criteria they find most important since they are dealing with applications which often change and these changes need to be applied rapidly. Customisability and modifiability will be the two criteria on which we will focus; these criteria will be evaluated by using feature description diagrams. This has been done because both customisability and modifiability are changes in the features or the features behaviour. We will define the difference between customisability and modifiability and propose a method for comparison based primarily upon feature description diagrams. The diagrams will be complemented with metrics extracted from the source code for each framework. The results indicate that the Spring framework holds a higher customisability compared to Spring. However, Blueprints holds a slightly higher modifiability than Spring. This indicates a slight advantage for Spring for the current comparison. Nevertheless, in order to completely evaluate both frameworks, more domains need to be compared. The contribution of this research is threefold; we provide a theoretical model for frameworks in general. Secondly, we provide a method for comparing frameworks based upon this theoretical model. Thirdly, we propose a method for comparing customisability and modifiability based upon features diagrams and metrics. We state that if the change of framework is made, a clear set of features will need to be defined which are allowed to be used and which are not. This due to the fact that Blueprints is more a framework of limitations and thus limiting complexity. Spring however is a framework of possibilities.
Subscribe to view the full document.
Anton Gerdessen, Master thesis Page 3 of 78 Acknowledgements This thesis is the result of my graduation project for the title of Master of Science at the university of Amsterdam. This thesis has been shaped by the influence of several people. First my university supervisor, Jurgen Vinju, who was always available for questions or discussion about the thesis. Secondly, Everest, who offered me the opportunity to perform this research. Also the supervisor at Everest itself, Onno de Groote and Mark Mastop, who provided me with suggestions and insight in the more practical issues for this thesis.