-
G. Westendorp, "A formula for the N-circumsphere of an N-simplex", https://westy31.home.xs4all.nl/Circumsphere/ncircumsphere.htm, April 2013.
-
Gerard Westendorp, "Space-time triangles", https://westy31.home.xs4all.nl/SpaceTimeTriangles/Space_Time_Triangles.html.
-
Charles G. Gunn, "Course notes: Geometric Algebra for Computer Graphics", SIGGRAPH 2019, https://bivector.net/PROJECTIVE_GEOMETRIC_ALGEBRA.pdf.
-
Anil N. Hirani, Kaushik Kalyanaraman, Evan B. VanderZee, "Delaunay Hodge Star", arXiv:1204.0747v4 [cs.CG]: Delaunay hodge, conditions for well-centred meshes.
-
Volker Springel, "E pur si muove: Galiliean-invariant cosmological hydrodynamical simulations on a moving mesh", arXiv:0901.4107 [astro-ph.CO].
-
Michael Reed, "Differential geometric algebra with Leibniz and Grassmann", https://crucialflow.com/grassmann-juliacon-2019.pdf.
-
WriteVTK.jl
... and Glitterati.
L u == ρ
B d u == 0
d u - f == 0
δ f == ρ
B f == 0
Can we rewrite this without δ or ⋆?
-
DA: need to remove harmonic forms from f: P(1-H) f. only for strong form?
-
DA: R[u]=0: von Neumann bc R[u]=D-1: Dirichlet bc
-
DA: mixed weak formulations have no hodge dual
-
magnetic bc: R=1 (or R=2?)
-
electric bc: R=2 (or R=1?)
-
DA: Hodge Laplacian is always well posed! choose complexes (and respective basis functions), then solve in the discrete with the same mixed weak formulation.
-
DA: trace is projection onto boundary; trace maps D-dim form onto (D-1)-dim form
-
DA: 0-forms: naturally piecewise continuous (FE), polynomial D-forms: naturally piecewise discontinuous (DG), polynomial - DOFs need to be located on either of vertices, edges, faces, etc. - must be unisolvent (be a basis?)