Error-Correcting Codes for Byte-Organized Arithmetic Processors
Peter G. Neumann, T.R.N. Rao · IEEE Transactions on Computers · 1975
This paper considers codes with radix r > 2 which are capable of correcting arbitrary arithmetic errors in any radix r digit. If each radix r digit represents a byte of b binary digits (e.g., r = 2b), these codes correct any combination of errors occurring in the b binary digits of any single byte. A theoretical basis for these codes is presented, along with practical considerations regarding their applicability.