Syntax Based Modularization: Invasive or Not?

Assume that the meaning of module A is modi ed by module B. If the modi cation can be detected by inspecting module A, then we say that the modi cation is invasive. Non-invasive modi cation makes it possible to reuse code which is \almost right" for a given purpose, even if we cannot change it. How much of a problem is it, if points of variability in source code must be marked explicitly? This paper presents a syntax based modularization mechanism where points of variability are always explicitly marked, and argues that this is a reasonable trade-o between safety and exibility.