ALGORITHM FOR EVALUATING THE QUALITY OF SOFTWARE PRODUCTS

An algorithm for evaluating the quality of software products is proposed, which includes two levels. At the first level, indicators are considered: operational efficiency, memory efficiency, error tolerance, unified communication procedures, data uniformity, traceability, consistency, information input and output speed, robustness, accuracy, access control, redistribution, availability, simplicity, information input and output volume, community, independence from other software, the availability of measuring tools, modularity, machine independence, brevity, access control, communication, informational content, completeness, the possibility of training. Applying the convolution algorithm to indicators of the first level, we can obtain indicators of the second level: efficiency, integrity, ease of use, practicality, evaluation, reliability, mobility, correctness, flexibility, the ability to use in other conditions, the possibility of interaction, security. Applying the convolution algorithm to the indicators of the second level, we can evaluate the quality of software. The convolution algorithm is the sum of indicators with weighting factors. The difference from other algorithms is that the coefficients are not calculated. Coefficients are subject to conditions. Firstly, the discreteness condition, i.e. coefficients can take only discrete values with a predetermined step. Secondly, the normalization condition, i.e. the sum of all the coefficients is equal to one. Thirdly, the priority condition, i.e. all coefficients are arranged in increasing order. Using this algorithm, we obtain several convolution values. The final result of the convolution will be their arithmetic mean. Using this algorithm, you can evaluate the quality of both software products and other products and services. In particular, this algorithm can be applied to assess the quality of educational services, medical services, etc.