From busyware to stableware

The integrated engineering of complex systems depends on, among other things, the supply of reliable, cost effective, stable components. However, as pointed out earlier by H.W. Lawson (1998), unnecessary complexity abounds in today´s computer and communication structures, including the platforms that host many complex, often critical systems. Thus, even if an organization accurately designs, develops, and supports reliable, cost effective, nonplatform components, deploying them in an unstable platform will counteract even the most stringent engineering and management efforts. The system will be no stronger than its weakest components (R. Colwell, 1998,). Societal burdens are already here. A recent example is the Windows NT error that caused the Yorktown, an Aegis guided missile cruiser, to be disabled. The Yorktown could not perform its mission and had to be towed into port. An otherwise well engineered system failed because of a platform bug. There are likely many platform related failures in a variety of critical application domains that are never made public. A concerted (hopefully industry driven) effort must be made to move toward the stable integrated platform solutions that Colwell indicates. This requires a transformation from what the author calls busyware to stableware