Војно дело (Jan 2016)
Assessment of the total software product development time
Abstract
The main motive of this paper on software engineering is to confront the engineering approach to the software development with the artistic one. Software engineering is the youngest engineering discipline and, as such, it still involves a high degree of arbitrariness. The very process of software development is largely lacking in the methods and techniques, as well as the relevant tools that would make it more efficient. The lack of development standards is even more prominent. However, the weakest point of software engineering is the great absence of proven and reliable measures and measurement techniques, as main tools for engineering verification of results. The importance of the problem is multiply emphasized by the need to provide high quality software and software-based systems. The aim of this paper is to develop and systematize an original formal procedure for assessing the development of information systems in the early stages of the software life cycle, by applying data model metrics. This objective was achieved using the experience of our engineers in the field of software development, as well as the results of other researchers who studied software engineering metrics. Data model metrics are calculated on the basis of data that can be read from the base data model, represented in the diagram of objects and connections (SPC), which is defined by four basic concepts: entities, connections, attributes of entities or connections, and values. The idea is to present the complexity of the process by the function of the number of these concepts and the number of attributes of entity types. Assessment techniques are the basis for successful planning and execution of software projects. This value is important because it serves as the basis for the calculation of other parameters essential for a software project, such as the costs or the total time of software product development. The authors used the statistical method, whose estimation procedures fall within the empirical parametric methods, although they have some properties of expert estimation methods. The developed assessment procedure is a step in the efforts to find suitable measures for assessing the size and complexity of the data model and estimating the costs and resources required for the development of information systems. Particular importance of the suggested metrics is that their values can be established at the earliest design stages of information systems. Also developed were metrics which, based on the knowledge of the data model, can quantify also the characteristics of the information system as a whole in the logic design stage. The suggested metrics were tested for specific models, and the obtained results are presented.
Keywords
