On Fri, Apr 16, 2010 at 5:28 PM, Alec Mihailovs
<[email protected]> wrote:
> On Apr 16, 5:05 pm, David Joyner <[email protected]> wrote:
>
>> Do you mean A is the matrix of all n^k codewords, but you remove the
>> columns corresponding to the erasure positions? I don't know what you
>> mean by A.
>
> I meant the submatrix of G - I used something like that in the decode
> procedure in another post, just replacing A.transpose() with a random
> matrix.
>


Perhaps I don't understand your program, but it appears to not address
the issue. Here is the algorithm, if I understand the question correctly:

Let I denote the subset of range(n) which represents the erasures.
Let v denote the vector in GF(q)^n which you want to decode.
Let C denote the [n,k,d] code with generator matrix G (so the
rows of G are a basis for the vector space C over GF(q)).
For each w in GF(q)^n, let w^I denote those coordinates of w not in I.
Let L = [] be an empty list.
For each c in C
  if c^I = v^I then append c to L
return L

This gives you the list of codewords desired.

I don't see how the output of your programs agree with this.



> I looked at the C.decode (for C in the original post), and it does
> something different - it corrects the errors. Perhaps, the better name
> for it would be C.correct ?
>
> Also, perhaps, just simple linsolve can be used (in my decode) if it
> exists in Sage and works for overdetermined systems of linear
> equations.
>
> Alec
>
> --
> To post to this group, send email to [email protected]
> To unsubscribe from this group, send email to 
> [email protected]
> For more options, visit this group at 
> http://groups.google.com/group/sage-support
> URL: http://www.sagemath.org
>

-- 
To post to this group, send email to [email protected]
To unsubscribe from this group, send email to 
[email protected]
For more options, visit this group at 
http://groups.google.com/group/sage-support
URL: http://www.sagemath.org

Reply via email to