The paper Nonisomorphic Verdier Octahedra on the same Base, by Matthias Künzer, popped up on the Arxiv today. It answers the following question: Given a morphism $f : A \to B$ in a triangulated category the cone on $f$ is determined up to a (non-unique) isomorphism; is the same true for a Verdier octahedron constructed from a commutative triangle? The answer is: No, there exist non-isomorphic Verdier octahedra constructed from the same base! The non-uniqueness mentioned above has been a thorn in the side of the standard axioms for triangulated categories and this result doesn’t exactly make their life easier!

To prove this result Künzer uses something called Heller triangulated categories which he describes in detail in the paper Heller Triangulated Categories. These two papers are definitely being added to my list of things that I want to understand!

On a related note, J.P. May has proved that the standard axioms for triangulated categories are over-determined in the sense that one of the axioms can be proved using the other three. This is explained in his paper The Axioms for Triangulated Categories.