I want to used a concatenated coding scheme. An inner code that can correct some errors and also detect errors (it therefore mark them as erasures for the outer code) . The outer code can correct the errors and erasures
This is not possible because if a code detect errors, it might not fully know where all these errors do occur. Error correction is one process, error detection is another process where for example, any linear error-correcting block-code can detect up to 2T errors but only correct up to T errors.
CRC are special block detection coding but you cannot detect where the errors are. Again detecting error is one topic, knowing where the error occur then becomes correcting. You need more redundancy to correct than to detect !
thank you for the answers Mr Albert, allow me to raise two points:
it is true that knowing the position of the errors is equivalent to correcting ...only for binary codes.
the reed solomon codes, for example, often use erasure information in their decoding process. how can we provide this erasure information, that is my question !