Newsgroups: sci.electronics,comp.dsp,sci.image.processing
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!ix.netcom.com!netcom.com!mejohnsn
From: mejohnsn@netcom.com (Matthew Johnson)
Subject: Re: Error-control coding
Message-ID: <mejohnsnD4pHG4.8Hx@netcom.com>
Organization: NETCOM On-line Communication Services (408 261-4700 guest)
References: <3iuca2$ig8@charnel.ecst.CSUChico.EDU>
Date: Tue, 28 Feb 1995 10:36:04 GMT
Lines: 17
Sender: mejohnsn@netcom19.netcom.com
Xref: glinda.oz.cs.cmu.edu sci.electronics:121642 comp.dsp:17213 sci.image.processing:12978

Surely one of the sources you mentioned must include the information
allowing you to compute the maximum number of errors correctable given
the parameters of the code.  I would do so myself if I had one of my sources
handy, since I canot remember the formula myself either.  In Berlekamp's book,
which was my favorite reference on the topic, the information was concealed
in the introductory chapter, the chapter on t-error correcting codes and the
chapter where R-S codes were introduced as a special case of the BCH codes.
In any case you must expect that if there are more than t errors in the block
you cannot correct them.  There is some supplement to the syndrome that you can compute to generate a "cannot correct" flag if there are not TOOmany more
than t errors, but even that fails if there are simply too many errors!
It sounds to me like your code is probably working just fine, but you are
giving it too many errors, but since I don't have the references in front of
me, you must double check yourself.

Matt Johnson
mejohnsn@netcom.com

