Transforming natural language into controlled language in requirements elicitation: a pre-conceptual-schema-based conceptual synthesis

Abstract

During the requirements elicitation process, a group of analysts and stakeholders identify, capture, and integrate requirements. Textual or graphic descriptions capturing the most relevant concepts from the domain of a software application development are generated. Commonly, the initial phases—identifying and capturing requirements expressed in natural language—are executed by using techniques in which high analyst intervention and comprehensive knowledge of the context and the problem domain are required. Thus, a subjective, ambiguous, and error-prone process is implied, causing losses in the generation of the initial domain models (specified in a controlled language). In this paper we provide a synthesis of trends and conceptual approaches found in the state of the art concerning the natural language transformation into controlled language during the requirements elicitation process. Finally, we propose a pre-conceptual schema for representing the conceptual framework of the transformation process.