Beyond the Wikipedia definition, I've heard the term used to describe the phenomenon of people looking at an ugly codebase with tons of special cases and expectations, thinking they can create a leaner version only to end up going nowhere because all those special cases and expectations that made for ugly UML diagrams had reasons to exist and the overconfidence led to feature creep to justify the time investment to stakeholders. I.e. a good explanation for why it's often better to refactor than rewrite.