Answers are grounded in this paper and cite their sources.
We give a smooth counterexample to the mesh-independent second-order difference-quotient estimate in Becker and Rannacher's 2005 Oberwolfach report, equation (5), represented by UnsolvedMath record OWR-1058-002. The fixed polynomial u(x,y)=xy(1-x^2)(1-y^2) on the unit square has homogeneous Dirichlet data. Its conforming piecewise affine Ritz projections are computed on uniformly shape-regular meshes whose maximum diameter tends to zero. They converge in the energy norm, but an interior normal-derivative jump, scaled by the edge length to represent a second-order difference, is unbounded. Exact local stiffness matrices yield a positive tridiagonal recurrence and an explicit geometric lower bound. A separate fixed-outer-mesh family is nested. The five-page English manuscript includes the mesh construction, complete proof, scope qualifications, a vector figure, and reproducible exact-arithmetic and numerical diagnostics. The separate averaging conjecture and specified adaptive refinement rules are not addressed. This is an unrefereed preprint, with no claim of formal verification or absolute priority.
No takes yet. Share an insight, caveat, or question.
Alper Ferudun (2026) studied this question.
Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context: