An architecture for defining features and exploring interactions

The last decade has seen an explosive growth in the development of telephony features. The description and design of new features are fraught with errors due to this growth’s impact on our ability to recognize interactions and the current practice of describing a feature’s requirements using natural language. While the use of natural language eases the communication of requirements between the designer, customer, and developer, it introduces the potentially fatal flaw of ambiguity. Additionally, these requirements documents are rarely updated to reflect interactions with newly developed features. This paper presents an overview of a natural language-based system currently in development that converts English-based telephony requirements into a knowledge-based representation. The goals of this conversion are to create an unambiguous understanding of the requirements of the described telephony feature, to create less ambiguous written requirements documents, to automatically update existing requirements documents when new interactions are found, and to assist the feature designer in locating potential interactions with other features.