Abstract: A method for generating correction codes that can correct a single error and detect two adjacent errors in an information word for a range of (n, k) values, where k denotes the length of the information word in bits and n denotes the length of the coded information word in bits. The method generates a parity check matrix, then multiplies a received word by the parity check matrix to produce a syndrome corresponding to one of two mutually exclusive sets of syndromes if the word contains at least one error. Information in the word is corrected by inverting a bit containing an error if the produced syndrome corresponds to one of the sets of syndromes. An uncorrectable two bit adjacent error is reported if the produced syndrome corresponds to the other of the two sets of syndromes and no error is reported if the produced syndrome contains all zeros.

Abstract: An error correction and detection technique provides a correction code for correcting single bit errors as well as detecting but not correcting two adjacent bits in error. A received word, which may contain errors, is multiplied by a parity check matrix to produce a syndrome corresponding to one of first and second mutually exclusive sets of syndromes if the received word contains at least one error, each single bit error in the received bit word corresponding one-to-one with a member of the first of the sets of syndromes and each two bit adjacent error corresponding non-uniquely to a member of the second of the sets of syndromes. A syndrome containing all zeros is produced if the received word contains no errors. One bit data errors in the received word are corrected, two bit errors are reported, and no action is taken if the word contains no errors.