Autili et al. describe the EAGLE approach for software integration where the integrator does not know the boundary conditions for a system composed of working components. They claim that the reuse of working software demands that our software production processes radically change their focus. In addition, from the “more traditional aspects of efficiency and resource optimization, to adaptability and fit-for-purpose, [...] inefficient usage of resources and redundancy” must be accepted.
This paper is not an easy read, but seems to suggest that software system realization will become a trial-and-error approach of finding working components, hooking them together, checking to make sure they perform essential features, and accepting certain inefficiencies and uncertainties. Perhaps the execution space should be bounded by an encapsulation program, but this technique is not mentioned in the paper.
Figures 1 and 2 give insight to the future of software realization and require careful study. As we reuse more working software, we will need ways to determine how well the integrated components meet our needs.
If you find yourself doing more integration than development, study this paper carefully for it may forecast your future work. Nevertheless, it could benefit from a good edit.