SCAT: An Automatic-Programming Tool for Telecommunications Software

Engineering

Volume

Proceedings of the AAAI Conference on Artificial Intelligence, 5

Applications

The size, complexity and long life-time of telecommunications software, e.g. the programs for store program control (SPC) telephone exchanges, call for an increased software productivity and maintainability other than an improved quality. The availability of programming support environments based on standardized specification and programming languages greatly improves the software development process. Artificial Intelligence techniques are very promising aiming at further improvements and can provide a short-term payoff especially within an evolutionary approach leading up to an hybrid programming environment, i.e. a software environment made of both conventional and intelligent tools. The paper describes an intelligent tool, dubbed SCAT, based on ideas exploited by various automatic programming systems, like CHI, Programmer’s Apprentice and DEDALUS. SCAT is strictly related to the telecommunications domain, thus it differs from other systems in the domain specifity. SCAT partly automatizes the most crucial phase in the software development process, i.e. the transition from project’s detailed specification to the actual software implementation. SCAT has been tested in a few experimental software developments and in an actual application,i.e. the message handling system (MHS) to be made available in the Italian public packet switching network (ITAPAC).