This dissertation deals with the dynamics of collective action in free and open source software (FOSS) projects. Thereby the following questions are primarily answered:
i) Is the problem of collective action shaped by dynamical forces?
ii) How are different types of motivation related to the dynamics of collective action?
It is basically hypothesized that different development phases provide distinct incentives, opportunities, and limitations for contribution to FOSS. It is therefore assumed that developers with different types of motivation are differently attracted by diverse development phases. The relationship between development phases and motivation is firstly modulated by network externalities and switching costs respectively by the S-shaped production function of software. Secondly, the relationship is heavily affected by incremental decision making that changes payoffs for the developers. It improves the information of the successor and allows antecessors to make a credible commitment and fosters conditional cooperation.
The conducted empirica1 study supports the basic assumption that there is a relationship between development phases and motivation. Based on the findings of this dissertation one may presume that i) collective action is, depending on the production function of a collective good, to a greater or lesser extent shaped by dynamical forces and that ii) different types of motivation are related to the dynamics of collective action a1though there is no fix or universa1 relationship for a11 types of collective action.
As hypothesized the motivation related to human capital increases in the initia1 phase, decreases in the intermediate phase, and increases in the mature phase of FOSS development. The relationship between peer recognition and FOSS development is, as suggested, u-shaped. Investments in reputation increases with FOSS development going on not logarithmically as suggested but linearly.
However, not all predictions concerning the relationship between motivation and development phases proved to be correct (or almost correct). Contrary to the hypothesis pro socia1 motivation increases about linearly during development. Also not in line with the hypothesis personal requirement answers to human capital.
Fun seeking was wrongly thought to decrease linearly in FOSS development because autonomy was predicted to vanish linearly. The data shows, however, clearly that fun seeking remains almost constant. The reason for this rather surprising finding may be that autonomy remains a1so more or less stable in all development phases. This result may be connected to the self-selection mechanism found in FOSS development. Because fun seeking is the most important type of motivation in FOSS development this has far reaching implications.
Commercia1 projects may have a hard time to reproduce self-selection mechanisms. Employed programmers should select tasks to accomplish that are in the interest of the employer. In consequence, commercial projects that intend taking advantage of the FOSS production model should either deliberately foster other motivation types than fun seeking or treat the autonomy of their employees with special care.