Don't get me wrong I have nothing against doing that theoretical upper bound research. The drama is that I struggle to find real improvements to maxflow (or more generally flow?) algorithms that came from this theoretical progress. Isn't that one role that theoretical algorithmic should fill?
