4 Wrong factor on B coefficient (single coeff version)
5 -----------------------------
6 (Probably fixed in the "calculator object" versions!)
7 Under Kristensson normalisation (with CS = -1), my code gives
8 B(1,0,n,n)/B(1,0,n,-n) == -(2n)! at (x,y,z) = (x,0,0)
9 (expected plus or minus 1).
10 A-coefficients seem to behave correctly.
11 BUT THE CSPHASE MIGHT BE WRONG ON BOTH SINGLE A AND B (TODO TEST)
13 Xu's antinormalisation
14 ----------------------
15 "Xu's antinormalisation" is broken (most likely in legendre.c and maybe
16 also in qpms_types.h) – the plane wave test fails and the spherical wave
17 reconstruction as well (but the translation coefficients match the
22 Plane wave decompositions gives wrong value on the longitudinal part.
23 The implementation of the L coefficients OR the longitudinal waves
24 is thus probably wrong.
26 Scattering result asymmetry
27 ---------------------------
28 The lattice scattering code (such as finitesqlatzsym-scattery.py) produces
29 asymmetric results where one should not get them, due to the system symmetry.
31 It seems that the asymmetry appears mostly in the y-direction (i.e.
32 for example the scattering/absorption cross section at k = (kx, ky, kz)
33 is not exactly the same as k = (kx, -ky, kz).
35 What has been checked (hopefully):
36 - The flip operators for electric waves
37 - Some weird kind of rounding or other numerical error depending on
38 the position ordering of the matrix (randomized indices give
41 What has not been checked (so well):
42 - The x,y,z-flip operators for magnetic waves (i.e. are they really
43 supposet to bejust the
44 same as the operators for electric waves, just with opposite sign?)
46 - the translation operators
49 Singular value asymmetry
50 ------------------------
51 Similar as the scattering result asymmetry, although not necessarily
52 only in the y-direction?
56 Assertion failed in gaunt_xu for test_vswf_translations.c and high values of LMAX
59 Moreover, the non-normalized legendre functions that are used in translations.c
60 are likely to overflow for even smaller values of l.
61 Therefore TODO: modify translations.c in a way that uses normalised legendre functions everywhere.
65 When compiled with optimizations, I get stack smashing error in qpms_trans_calculator_get_2DFT_longrange_AB_arrays().