Growth and Change Dynamics in Open Source Software Systems

摘要: In this thesis we address the problem of identifying where, in successful software systems, maintenance effort tends to be devoted. By examining a larger data set open source systems show that is, general, spent on addition new classes. Interestingly, efforts base code stable classes will make those less as they need modified meet needs clients. This advances state art terms our understanding how evolving grow and change. We propose an innovative method better understand growth dynamics systems. Rather than relying commonly used analysing aggregate system size over time, analyze probability distribution range metrics change time. Using approach find process evolution typically drives popular within gain additional clients time increase popularity makes these change-prone. Furthermore, once have been released, resist modifications do undergo are small adaptations rather substantive rework. The methods developed can detect releases with systemic architectural changes well identify presence machine generated code. Finally, also extend body knowledge respect validation Laws Software Evolution postulated by Lehman. consistent support for applicability following laws evolution: first law Continuing Change, third Self Regulation, fifth Conservation Familiarity, sixth Growth. However, analysis was unable evidence other laws.