THE PROCEDURES AND OPERATIONS USED IN DETERMINING THE QUALITY OF SOFTWARE

Abstract

The rapid development of information technology has led to the development of methods and tools to build software of varying complexity and for different applications. Increased confidence in the software and their ability to perform more complex functions increases, in turn, is responsible for performing these functions. This increases the interest of developers and users of software to analyze their quality, to assess the level which is necessary to apply the procedure of measuring, monitoring, evaluation, comparison and preference. This allows not only to identify and predict the level of quality of software, but also to study the characteristics of software depending on various factors. Quality of software is the totality of features and characteristics of the software that would cause their ability to meet the specified needs. Most of these properties and characteristics are ordinal values displayed on the scales of the order of or ordinal scales.A common feature of all known nomenclatures software quality indicators is that they are multi-level hierarchical structure and perform quality assessment indicators going from a lower level to the upper structure. When determining the individual indicators of quality of software that are on the lower level of the hierarchy of the quality system, the procedures of measurement and evaluation. Measurement is an empirical and objective procedure which preceded the establishment of clear concepts and properties of the corresponding scale. Measurement is homomorphism reflection properties of the object. But there are a number of properties that characterize the quality of software for which the measurement is problematic (eg, utility, meet certain requirements, etc).For non-physical properties and those that can not be observed directly we use the procedure of evaluation. Evaluation is empirical and subjective. In terms of the representative theory evaluation is a subjective homomorphism. Improving of the evaluation is to reduce this subjectivity. One of the directions of this improvement is the use of verbal and numerical scales.In determining the weight coefficients for the aggregation of individual quality indicators used operation preference. In terms of the representative of the theory it is based on the endogenous choice and is subjective. Therefore, in determining the weight coefficients of some parameters are used as a measure of consistency of expert decisions. In determining the weighting coefficients estimates are used that are adequate statistical measures for ordinal variables, ie the median and more effective assessment - median Walsh.Thus, the bulk of the properties of software is an ordinal variable, then there are certain requirements for operators of collection (aggregation). In the case where all weights equal to one quality indicators, aggregation operators can use the median, T-norm and T-Connor, emulator arithmetic mean OWA. When using verbal and numerical scales using aggregation operators for numerical and verbal scales. But it is a problem of estimation convergence results aggregation.