Assessment and comparison of local and global SW engineering practices in a classroom setting

Abstract

In this paper we present assessment and comparison of local and global software (SW) engineering practices based on our software engineering (SE) class jointly taught for the last there years between San Francisco State University (SFSU) and the University of Applied Sciences, Fulda University, Germany. We define global SE practices as those used when the team members are for the most part distributed in terms of location and time, and hence do not meet regularly in person. While global SE practices have become a significant mode of SW development, surprisingly little formal measurements and comparisons have been done to understand it and compare it with traditional (local) SE practices. Moreover, the challenge remains as to how to effectively prepare students and employees for this environment. The key contribution of this paper is an attempt to objectively compare differences between local and global SW engineering practices from the developers' perspective, using measurements from our SFSU/Fulda SW engineering class as a simulation of a real-world environment. We also propose practical SE teamwork assessment methods based on our comparison measurements. In analyzing the differences between local and global SE practices we measure and analyze the following five main factors: quality of final delivery (including development of milestone documentation), progress impediment factors, expended effort, level of collaborative activity and teamwork problems. One of the surprising findings is that in today environment the differences between local and global SE practices are blurring, since local groups have to employ many practices of global groups, e.g. are teams of diverse cultures who seldom meet each other in person. We also confirm that global groups spend significantly more effort in producing comparable deliverables.