Talk:Scattering-matrix method

Inaccuracy

There are lots of methods to solve Maxwell's equations that involve computing scattering matrices, involving many different bases. There are eigenmode-expansion methods (e.g. camfr is one example), there are methods using spherical harmonic bases, there are scattering-matrix computations using boundary-element methods, there are methods using planewave expansions....

To describe as "the" scattering-matrix method a technique restricted to N cylinders, or to using cylindrical basis functions in general, reflects such a fantastically limited understanding of numerical scattering methods as to be grossly misleading.

(Also, to claim that it doesn't give "spurious effects" like "other methods" is misleading. Any convergent numerical method, including FDTD, produces the correct result (e.g. the L2 error goes to zero) in the limit of increasing resolution (or number of basis functions). Scattering-matrix techniques are no different: in general, they are convergent-series expansions of the solution in some basis, that produce the correct result in the limit of infinitely many basis functions, but of course for any finite truncation of the basis they have some errors. All convergent numerical methods are derived from "first principles"....scattering-matrix methods are no different in this regard.)

— Steven G. Johnson (talk) 16:17, 30 September 2009 (UTC)

Hence, you think this article should be discarded then?

--Hyprwfrcp (talk) 14:59, 27 May 2019 (UTC)