The study of discrete analogues of concepts from differential geometry: definitions of curvature, connections, geodesics, and variational structures on discrete objects (meshes, graphs, simplicial complexes) designed to reflect and approximate smooth geometric behaviour.

Principle

DDG is governed by the principle of structure-preserving discretization: discrete models should preserve key geometric identities, variational principles, and invariants (e.g., Gauss–Bonnet, Stokes-type theorems) and converge to smooth notions in appropriate refinement limits.

Demonstration

A standard demonstration is the discrete Gauss curvature assigned to vertices of a triangulated surface via angle deficits; summing these deficits reproduces the Euler characteristic in exact discrete form, mirroring the smooth Gauss–Bonnet theorem.

Misapplication

Treating any arbitrary mesh or finite-difference scheme as DDG without ensuring preservation of geometric identities or convergence properties; for instance, using naive discretizations that destroy discrete curvature invariants and break conservation laws.

Consequence

DDG yields computationally effective geometric algorithms for processing surfaces, simulation, and graphics, provides provable convergence results towards smooth geometry under refinement, and offers combinatorial formulations of classical theorems useful in numerics and topology.

Reversal

The reversal is classical smooth differential geometry or purely combinatorial graph theory without geometric compatibility: continuous theory emphasizes infinite differentiability, while combinatorial theory may ignore metric or curvature structure.

Boundary

Applies to finite combinatorial complexes and meshes that carry geometric data and notions of discrete differentiation; excludes purely statistical network analyses or graph algorithms that lack geometric interpretation and schemes that do not aim to preserve differential identities.

Semantic Tension

Tension exists between combinatorial discretizations that are easy to compute but nonconvergent and structure-preserving discretizations that are more constrained; DDG privileges preservation of geometric structure often at some computational cost.

Synthesis

Discrete Differential Geometry constructs finite, combinatorial models of smooth geometric concepts that preserve key identities and variational structure and converge to their smooth counterparts, balancing computational tractability with fidelity to differential geometry.