Logical network for stored data failure correction - uses redundant code to correct bit failures in integrated circuit network

摘要

A logic network for correcting bit failures in stored binary data uses only one decoding network. The information is held in a register in a highly redundant code which is used to detect and correct any errors when the data is read out. A closed loop network with suitable logic properties reconstitutes the data correctly. With a sufficiently redundant code, the information may be corrected even if several bits have become corrupted. A suitable code for this purpose is briefly described, and consists of digital words built up from test characters. A voting system is used so that faulty readings may be rejected.