Рамка за развоj на научни апликации

Abstract

The constant progress in the area of natural sciences, undoubtedly imposes the need for the development of scientific applications that simulate natural phenomena. The lack of using software development models and software engineering practices during the development process of software applications results in poor quality, problems in the development process and later problems with the code refactoring, optimization and maintenance. Software Engineering offers a great set of practices for each stage of the development process that need to be followed in order create applications according to generally accepted standards for quality software. As a result of the specificity of this type of applications, the standard software engineering practices cannot be fully implemented, but appropriate modifications and adjustments to existing software development processes must be proposed. In this doctoral thesis a framework for development of scientific applications is proposed including: scientific applications’ development model and scientific applications’ quality model. The application development model defines the process of developing applications in stages, while the quality model gives opportunity to assess the quality of scientific application through attributes and metrics. The main purpose of the framework is to increase the quality of scientific applications and to change current scientists’ practices. This doctoral thesis also presents scientific application modeling and developing techniques that include applying of the different approaches to development and programming models, requirements specification, software testing, etc. The application of this framework is confirmed by a practical example, and additionally a comparative analysis for evaluating the quality of two scientific applications by using the proposed quality model is presented. The use of the framework in the development of scientific applications results in a development process organized in phases, specified requirements, defined test cases, understandable code, automatic testing, quality assessment of applications according to already accepted quality standards, etc.