A Class of Error-Detecting Codes with Absent Parity Bits

All possible data symbols are divided into two categories, designated
A
and
B
. Transmitted data is subject to the constraint that all information appears in fixed locations within any message. In addition, the message is divided into segments, with each segment containing symbols of only one category. In all messages each segment is identical with regard to the category of data contained within that segment. Subject to these constraints, a class of error-detecting codes with absent parity bits is derived. The derived class is capable of true error detection of single-bit errors since the addition of one parity bit per symbol does not improve the probability of detecting such errors. Suprisingly, codes subject to the initial constraint can find fairly wide application in weather data transmission equipment, airline reservation systems, commercial data communication networks, and other fixed-format type information-handling and processing systems. The paper concludes with an introduction of parity bit considerations subject to the initial constraint and indicates that a single parity bit will detect the presence of two random single-bit errors per symbol in 50 per cent of all cases.