> One could imagine an IDE that used a database derived by slurping in text > files and spitting them back out in some well-defined and reproducible > form for interaction with a revision control system. Isn't this how IBM's VisualAge works? It is not file-based, either, and so faces the same issues. -- O.L.