Obstruction theory in model categories. (Q1421911)

Given a commutative square made up of \(f: A \rightarrow X\), \(p : X \rightarrow Y\), \(i : A \rightarrow B\) \(g:B \rightarrow Y\) with \(p \cdot f = g \cdot i\), the standard extension/lifting problem looks for a map \(B \rightarrow X\) which makes the triangles commute. Classical obstruction theory is a detection principle for existence and uniqueness of such maps. The authors generalize this to a proper, pointed, model category. They say that \(i : A \rightarrow B\) has an obstruction theory if there is an object \(W\), and there is a weak homotopy class from \(W\) to the fiber of \(P\) so that the lift \(B \rightarrow X\) exists if and only if the map from \(W\) to the fiber is weakly null-homotopic. They also require such an obstruction class to be functional in \(p\). The main results of the paper, in section 7, classify cofibrations which have obstruction theories with respect to all fibrations. As expected, there are many technical points, but the paper is well-written and includes, with the last section, many interesting examples.