Joseph McManus, AT&T Network Systems and Teresa Garland, Inference Corporation
Over time, the inflexibility of a large software system can inhibit the growth of the business it supports. This paper describes a situation encountered at AT&T’s Merrimack Valley Manufacturing Plant. The problem domain is inventory parts allocation in a complex environment. Two systems were developed on PCs to provide front and back end processing of customer order and inventory data stored in a mainframe. Data driven rules access the user data via terminal emulation software, and a rule-based program provides in-depth problem analysis. The systems have been used successfully for eighteen months.