| dc.description.abstract | This thesis investigates contractual incompleteness in software development projects. Specifications requirements and development effort estimates are traditionally at the core of software development contracts and it is this kind of contractual incompleteness I have investigated in a multiple case study of four Norwegian software development projects.
Removing specifications and estimates is a radical proposition; it challenges established notions of accountability, of investment and of control. It’s as such not a viable or feasible option in every context. Academic research on Contractual Incompleteness and my empirical research points towards it being a possible and valuable practice in cases where there is trust between the contracting parties, where there is considerable uncertainty, volatility and ambiguity. Theories on cognitive biases in addition tell us that our ability to predict and plan is limited.
Deliberate contractual incompleteness seems to be valuable to the projects in my study in the sense that they are able to grow their product, their software more organically, build and learn continuously. An early and small increment is used in order to understand what the real requirements are. But in the process of shedding some traditional planning practices, they also seem at risk of losing sight of the big picture. Practices for these kinds of projects then need to deal with both seeing the big picture and facilitating a more organic growth of software products, while at the same time keeping day-today development work lightweight and structured. There also needs to be developed more structured ways to build and ensure trust; both in contracting and within projects. | nb_NO |
