Experiments and measurements for systems integration
The engineering of sytems integration (type B) software requires new ways of experimental evaluation, because high enough fidelity of modeling is impractical due to complexity. Also, traditional software engineering practices are insufficient: software as well hardware components must be integrated, and experience in the application domain is also necessary. This paper outlines the problem but does not offer specific solutions. Rather, it intends to call the attention of the software engineering community to spend more effort in this area of research.