CHAPTER VII - The Place, If Any, of Coding Techniques in Failure-Tolerant Digital Networks

Abstract

The practicality of coding techniques for processing unreliable information depends upon whether or not the unreliable information is to be processed by reliable or unreliable circuits. When the information processing circuits have far greater reliability than their input signals, as for example in the decoding of redundant communications or in the decoding of redundant bits stored in magnetic tape or core memories, then coding can be practical and it has been used. This chapter discusses the practicality of using coding in computers when failures can occur in the digital logic. The qualitative aspects of using coding theory that appear to cause impracticality are clearly developed, but no quantitative treatment has been found. The topic of using coding theory in computers for component failure correction has had a long and somewhat troubled history. The complexity of the decoding operation is the crucial limitation on the usefulness of coding techniques in computers.