On Sat, Dec 12, 2009 at 10:35 AM, Gerard Bosch <[email protected]> wrote:
> Hello,
>
> First of all I have to say that I'm newbie with Sage. And I read that
> it's a good idea to start a disscusion here in order to implement new
> functionalities in Sage.
>
> I'm working with the implementation of the Golay codes, in particular
> for the binary Golays (G23 and G24). I have made a new python module
> for this purpose with an specific decoder method. The algorithm I have
> implemented is based in the weight of the syndromes of the codewords
> to decode. The algorithm is an incomplete decoding scheme (i.e. can


Syndrome decoding. Is your procedure faster than the syndrome
decoding already implemented in Saeg (see the sage/coding/decoder.py
module for details)?


> correct a number of errors <= correcting capacity (t) of the code; 3
> in the case of Golays). So is capable to correct up to 3 errors (and
> detect up to 4 errors simultaneously for G24).
>
> I already have implemented it and works fine and quick to decode up to
> 3 errors. But now there is only a problem and it's that I don't know
> how to treat the no-decoding cases (i.e. the cases in G24 which the
> word to decode contains 4 errors). Now the method returns a vector
> with the corrected word (if can decode) or the string "*" if more than
> 3 errors was found (only for G24; G23 is perfect).
>
> I think it's a problem that a function that should return a codeword
> could return a string ( "*" ) to comunicate that decoding is not
> possible, so I would like to know how should I treat this cases.
> ¿...raising an exception or something like this?, how the error cases
> are usually treated in Sage?


My opinion is to either return

(a) an arbitrarily selected (but deterministic) closest codeword,
(b) revert to list decoding,
(c) both - namely return by default a closest codeword, but allow
a keyword option which returns the list of all closest codewords.


Maybe others have a different opinion though.

Thanks for working on this!!


>
>
> I hope for your replies and any advice that could help me with this
> question or related to devel process (steps to follow, etc...)
>
> Thanks,
> Gerard Bosch.
>
> --
> To post to this group, send an email to [email protected]
> To unsubscribe from this group, send an email to 
> [email protected]
> For more options, visit this group at 
> http://groups.google.com/group/sage-devel
> URL: http://www.sagemath.org
>

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

Reply via email to