The right process for each context: objective evidence needed

The growing importance of software in ever more technical systems has led to new demands with respect to developing software. The demand for more functionality, higher quality, and faster delivery hence poses major challenges to the software industry. The software process community has responded with a variety of different development processes such as the waterfall model or the incremental commitment model, however, the number of late or failed projects has not decreased as much as it was desired. In the new millennium, agile development approaches promised a new way out of this dilemma. After several years of heated discussions, it is now time to evaluate applicability, advantages, and challenges of different software development approaches based on sound, empirical evidence instead of anecdotes and hearsay. This paper briefly investigates the major differences between agile and traditional approaches, illustrates the difficulties in selecting the "right" approach for a given project, and proposes hypotheses for empirical evaluation, in order to build a solid body of knowledge that can be used for said selection.