Schumaker’s lower-bound equality conjecture for C^1 cubic spline spaces

For every nondegenerate planar triangulation T\mathcal{T}, the dimension of the C1C^1 piecewise-cubic spline space satisfies dim⁡S31(T)=PT(1,3)\dim S_3^1(\mathcal{T})=P_{\mathcal{T}}(1,3), where PT(1,3)P_{\mathcal{T}}(1,3) is Schumaker's classical lower-bound expression, including the local correction σ\sigma for singular interior four-stars.

References

Primary source

arXiv

Progress summary

Refreshed
Claimed solved

A 2026 preprint claims an explicit counterexample, but the alleged disproof has not been independently verified.

The conjecture asks whether Schumaker’s lower bound always equals the dimension of planar C1C^1 cubic spline spaces, including the correction term σ\sigma. It arose from work of Strang and Schumaker and remained open for arbitrary triangulations.

Known results

  • Schumaker (1979, 1984) established the lower bound and the local correction for singular interior four-stars.
  • Billera (1988) proved equality for generic triangulations in the case (r,k)=(1,3)(r,k)=(1,3).
  • Schenck, Stillman, and Yuan (2019) established broad high-degree results and documented failures in other low-degree regimes.

2026 counterexample claim

The preprint Geometry-dependent rank defect in C1C^1 cubic spline space claims that an 1818-triangle family has dim⁡S31=34\dim S_3^1=34 at t=1/5t=1/5, while PT(1,3)=33P_{\mathcal T}(1,3)=33 and σ=0\sigma=0. If correct, this is a nonsingular global rank defect and disproves universal equality; the calculation remains unverified.

Current status (as of September 2026): The conjecture has not been verified as false; the claimed 1818-triangle counterexample remains unconfirmed, so the exact arbitrary-triangulation case is unresolved.

Sources

Solutions 0

No solutions have been posted yet.