METHOD FOR ANALYZING THE SOFTWARE REQUIREMENTS FOR FINDING THE VALUES OF QUALITY ATTRIBUTES

Abstract

The software market is growing very rapidly. Software is becoming more and more complex, and more and more responsibility is being placed on it. With the increasing reliability of hardware and the growing complexity of software, software quality is a growing concern for both developers and users, at least in terms of achieving business goals.

Currently, there is a contradiction between the growing complexity of software and the responsibility it carries, the expansion of software quality requirements, on the one hand, and the imperfection of models, methods and tools for predicting software quality at early stages of the life cycle, on the other. Therefore, predicting the level of software quality at the early stages of the life cycle based on quality attributes is an urgent task.

The developed structure of the software requirements specification, suitable for automatic processing by a system for predicting the quality level of software based on requirements, imposes certain restrictions on the formulation of requirements containing quality attributes and is intended for further preprocessing of software requirements specifications.

The method for analyzing the software requirements for finding the values of quality attributes provides the selection of values of software quality attributes from the natural language specification of software requirements, which can then be used to determine the values of software quality characteristics and for a comprehensive assessment of software quality. The developed method is important for full automation of requirements processing and complete elimination of subjective influence and human participation in the processes of information processing and knowledge acquisition. The method for analyzing the software requirements for finding the values of quality attributes is the theoretical basis for developing a module for automatic analysis of software requirements of the system proposed by the authors for predicting the level of software quality based on requirements.