The talk is about a search for geometric conditions (as opposed to obstruction theory) that imply trivialization of the tangent bundle for smooth H-spaces. The geometric conditions that I have studied all lead to classical examples