The vast amount of computation techniques for reputation systems proposed in the past has resulted in a need for a global online trust repository with reusable components. In order to increase the practical usability of such a repository, we propose a software framework that supports the user in selecting appropriate components and automatically combines them to a fully functional computation engine. On the one hand, this lets developers experiment with different concepts and move away from one single static computation engine. On the other hand, our software framework also enables an explorative trust evaluation through user interaction. In this way, we notably increase the transparency of reputation systems. To demonstrate the practical applicability of our proposal, we present realistic use cases and describe how it would be employed in these scenarios.