When approaching a new problem with FDR, it is tempting to dump a detailed description of the problem into CSP as quickly as possible and "see how well FDR copes." The results of such an exercise are frequently disappointing. The model produced is usually intractable, and ad-hoc attempts to simplify it tend to reduce its coherence and destroy any confidence in its accuracy. Much better results are achieved by starting from a minimal model and incrementally adding and testing features; in particular, incremental testing will reveal flaws soon after their addition to the model.
Note that even if the initial model is small, the choice of identifiers and careful commenting is still important. In particular, single character process and tag names are likely to cause readability problems.
Go to the Next or Previous section, the Detailed Contents, or the FS(E)L Home Page.