A way of encoding binary signals that aims to achieve the highest possible data transfer rate for a given signal frequency. The name is derived from the principle of operation, i.e. the signal line does not return to zero—make any transition—between a succession of 1 bits. The method was first used for communications signalling in which there was always a 1 bit at the start of a character and thus there was a predictable and acceptable short interval over which the sending and receiving devices had to maintain synchronism independently.

Many variants of the basic principle have been derived to overcome synchronization problems that occur at high speeds and long bit streams. See also disk format, tape format.