In this paper, we call to bridge the gap between what makes highly-concurrent optimistic algorithms work and current approaches for proving their correctness.