The Case for Electric Design of Real-Time Software

G. Booch (see ibid., vol.SE-12, no.2, p.211-21, Feb. 1986) has analyzed a problem involving the software of a set of free-floating buoys. The correspondence points out that Booch's analysis fails to address one important system issue, namely the fact that the software must support two concurrent activities, and shows that an analysis according to the M.A. Jackson method will reveal this difficulty at an early design stage. On the other hand, the Jackson approach does not deal with some configuration issues, which are handled in Booch's analysis. This shows that one method is sometimes not enough to address all important, systemwide aspects of a problem. Rather than arguing about which one design method is best, the author recommends taking an electric view and using any combination of approaches that yields important results in a given situation. >