The physics of General Relativity is deeply intertwined with the mathematics of Lorentzian differentiable manifolds. The latter provide excellent models of spacetime across a vast range of physical scales, encoding gravitational interactions into the curvature properties of smooth metric spaces. However, describing geometry in terms of the infinitesimal line element "ds" does not seem appropriate in the quantum regime near the Planck scale. -- Coming from different perspectives, geometers, classical relativists and quantum gravity researchers are actively investigating more general geometric settings, abandoning smoothness or even continuity of the metric space. The crucial question is what aspects of standard (pseudo-)Riemannian geometry survive, and what this tells us about the "essence" of curvature and geometry and about the ultimate nature of physical spacetime. Can we learn from each others' insights?