Ett ramverk för fördelning av arbetsinsats vid injektion och upptäckt av mjukvarufel
Sammanfattning: Software developers spend much time on finding and fixing software faults, both during the development and the maintanence of the system. Despite this, there hasn’t been much research done on the effort that these activities require. It is important, both for developers and clients of software systems, to know how they can use their resources as efficiently as possible. Since software errors can cause high costs and can result in serious consequences, it is therefore of interest to have a basis for how much time is spent on finding and fixing software errors. Software Fault Injection (SFI) is a method of injecting artificial bugs into software, which is used to assess a program’s reliability. This study looks at the possibility of using SFI to develop a framework in order to measure the effort in finding and fixing errors. An existing software system, EXIT, was injected with a set of predefined errors. They were then detected and corrected with activities obtained from previous research conducted by other researchers. The result was the Software Fault Injection/Detection Framework (SFIDF). A framework that can be used to measure the time distribution for injection, detection and the fixing of software errors.
HÄR KAN DU HÄMTA UPPSATSEN I FULLTEXT. (följ länken till nästa sida)