Abstract
Data communication has grown dramatically in the past two decades, both in technical sophistication and in usage. While channels and networks designed expressly for data use have emerged, the voiceband channels of the telephone network continue to be the major transmission medium; data sets, or modems, play a role analogous to that of the telephone in voice communication. Fundamental developments such as adaptive equalization along with bandwidth-conserving signal formats have allowed the modem to better match the characteristics of the analog channel, resulting in increased available throughput. The quest for further improvements, along with elegant implementations and attractive user features, provide a continuing challenge to communication engineers. When the need arose to communicate digital data at transmission rates substantially higher than telegraph speeds, first for defense about mid-century and then for industry in the late 1950's, all the ingredients required for successful implementation were at hand. Decades earlier, Nyquist had formulated the filtering or band shaping requirements to allow the independent transmission of a sequence of signal samples. Just prior to mid-century, Shannon had published his celebrated information theory, which showed engineers the maximum rate at which they could signal through a channel if only they were clever enough. The telephone network, especially in the United States, had reached a high state of development and widespread accessibility; it seemed like an ideal vehicle to carry the new data communication traffic.