Today, architectural component models are often used as a central view of architecture descriptions. So far, however, only a very few rigorous empirical studies relating to the use of component models in architectural descriptions of software systems have been conducted. In this paper, we present the results of a controlled experiment regarding the supportive effect of architectural component diagrams for design understandability. In particular, the goal of the experiment was to determine whether architectural component diagrams, provided in addition to a non-trivial software system’s source code, have a supportive effect on the ability of novice architects to answer design and architecture related questions about that system. Our study provides initial evidence that architectural component diagrams have a supportive effect for understanding the software design and architecture, if a direct link from the component diagram’s elements to the problem that requires understanding can be made. If such a direct link cannot be made, we found evidence that it should not be assumed that architectural component diagrams help in design understanding, for instance only by providing a big picture view or some general kind of orientation.

Top