readme

在matlab中

This code was written by Anand Sarwate for a class project at UC Berkeley
in the spring of 2005.  It has not been tested extensively, is slow, and
might break in some cases.

You can try and email asarwate at alum dot mit dot edu with questions,
but the code should be pretty self-explanatory if you understand the
relevant algorithms and some simple facts about Reed-Solomon code
representations.

In particular, there are two ways of representing Reed-Solomon codes -- 
as evaluations of polynomials at all the points of GF(2^m), and as 
a kind of non-binary cyclic code.  All of these algorithms use the former
as their starting point, but MATLAB's built-in RS encoder and decoder
use the latter (which puts the code in systematic form).  If you are
using the MATLAB functions to test the decoding, make sure to do the
appropriate conversions.

Function files included here:

gs_weighted -- performs the Koetter-Vardy algorithm using a multiplicity 
	matrix provided by the user.  This function is incredibly slow
	due to the interpolation step.  The root finding is pretty fast,
	however.
koetter_interp -- run Koetter's bivariate polynomial interpolation algorithm.
	This is awfully slow due to terrible MATLAB writing style, but a
	lot of Galois Field operations are not vectorized in MATLAB yet,
	so loops seemed like the way to go
koetter_vardy -- generate multiplicity matrix from the channel reliability
	matrix and a total desired number of interpolation points.
rr_factor
rr_dfs -- Ruckenstein-Roth root finding algorithm
multFromRx -- multiplicity matrix generation from hard-channel outputs
hasse_deriv1 -- take a 1d Hasse derivative
hasse_deriv2 -- take a 2d Hasse derivative
minwtdegree -- find minimum weighted degree polynomial