#1 some questions about code

[Anonymous]

usually when one talk about RS codes they use terms
like codes of (255,249) or (7,5,3) etc...

You use NPAR (meaning number of parity bits) how does
that translate into the formats above?

was there a specific reason why you used NPAR rather
than taking the normal RS code paramters?

Marcus

[Nobody/Anonymous]

Logged In: NO

First of all, I'm not a writer of this source but
I studied this code a long time ago by myself and want to
comment. NPAR here means 2t

Let's say RS code (n, k), then n-k = 2t
It means that if we want to send 'k' message, then we need
to add '2t' as a redundancy encoded by RS.

Jaewon Ahn (Chris)

[Nobody/Anonymous]

Logged In: NO

Convential notation is (n, k) k symbolizes message
symbols and n symbolizes total symbols. Therefore, NPAR =
2t = n-k
Your question above:
1) (255, 249) : 255-249 = 2t => t = 3 ( it means 3
errors can be corrected.
2) (7, 5, 3) is noted as (n, k, d)
where t = (d - 1) / 2
therefore t = 1 (any value below . wil be rounded
off)
This code is one-error correcting code.

I guess that author use the variable name 'NPAR' rather
than 2t because 2t is not valid name for a variable in C.

J. Ahn

[GANESH YELLAPU]

Hi,

In the example.c, erasure locations are added to the array erasures[nerasures++].
However erasures are not filled in the codeword[] before calculating the syndrome (inside decode_data())

In particular, the function byte_erasure() is not called to initialize the erasures in the codeword[].

kindly clear me in this regard.

Thank you,

Ganesh Yellapu