As we agreed on the past few days, defining precisely our decoder's types is not that easy.
I copy hereafter the definitions I wrote while working on trac #19897. As I removed this table from this ticket, I opened a dedicated ticket, trac #20001 We should discuss these definitions so we all agree on before putting them in Sage.
- always-succeed: Always return a closest codeword if the number of errors is up to the decoding radius
- complete: Can decode every word in the ambient space of the code
- dynamic: The list of types will be dynamically updated at construction time
- half-minimum-distance: Its decoding radius is at most half the minimum distance of the code
- hard-decision: The symbols of the input word are taken from a finite set of values without any clue on their reliability
- list-decoder: Returns a list of codewords
- might-fail: Might fail at returning anything at all
- might-error: Might return a codeword which is not a closest codeword
- unique: Returns a single codeword