Hands-On Design Patterns with C++

Acyclic Visitor

The Visitor pattern, as we have seen it so far, does what we wanted it to do. It separates the implementation of the algorithm from the object that is the data for the algorithm, and it allows us to select the correct implementation based on two run-time factors—the specific object type and the concrete operation we want to perform, both of which are selected from their corresponding class hierarchies. There is, however, a fly in the ointment—we wanted to reduce complexity and simplified the code maintenance, and we did, but now we have to maintain two parallel class hierarchies, the visitable objects, and the visitors, and the dependencies between the two are non-trivial. The worst part of these dependencies is that they ...