Reed-Solomon decoding: Euclidean vs Massey-Berlekamp

Hi There,

Is there a difference between the two algoririthms regarding
performance?  Do both the algorithms correct all correctable errors
and erasures (d_{min} >= 2s + e, s - errors, e - erasures)?

When would one use the one algorithm instead of the other?

Kind Regards,
Jaco

On Dec 20, 9:20 am, jaco.versf...@gmail.com wrote:
> Hi There,
>
> Is there a difference between the two algoririthms regarding
> performance?  Do both the algorithms correct all correctable errors
> and erasures (d_{min} >= 2s + e, s - errors, e - erasures)?
>
> When would one use the one algorithm instead of the other?
>
> Kind Regards,
> Jaco

The fact that the Reed-Solomon code can correct both errors and
erasures simultaneously has to do with the code being maximum distance
separable. Both algorithms can be used to the find error locator
polynomials but the Euclidean (Sugiyama-Kasahara-Hirasawa-Namekawa)
algorithm is much easier to understand, learn and teach, at least in
my experience. There are claims that the Berlekamp-Massey is somewhat
more computationally efficient but that may have to do with being much
older and most programmers are lazy to rework something that already
is doing a good enough job.