One of the more interesting approaches I've seen lately is in Andrei Alexandrescu's "An Algebra for Expressing Time Complexity of Generic Functions" article[1].

One of the more interesting approaches I've seen lately is in Andrei Alexandrescu's "An Algebra for Expressing Time Complexity of Generic Functions" article[1]. The gist is you (or more likely, the library writer) annotate functions with their complexity guarantees then you can make static assertions about your runtime performance expectations which result in a compile time error if the algorithms/data structures being used cannot meet them. In theory that would prevent the O(n^2) landmine. 1. http://erdani.com/d/bigo.html http://erdani.com/d/bigo.html