ITERATIVE APPROACH TO INCREASING QUALITY OF PROGRAMS TESTING

Abstract

The article deals with the problem of improving the quality of applied software (AS) testing. This testing begins with the pre-alpha version and ends at the stage of acceptance testing of the AS. This period includes the process of multiple execution of designed and developed system tests of AS, and includes additional analysis of the program in other ways. We refer to the testing process for a given period of time as a testing cycle (TC), which is usually implemented using the "black box" method without the tester's access to the program code. An analysis of existing methods of tests synthesis showed that these methods not always guarantee an acceptable completeness of testing. Therefore, in this paper we propose an approach to the development of additional tests, based on the analysis of the database of errors (DB) and the capabilities of programmers and testers. These additional tests should confirm or disprove the conclusions about the sufficient completeness of the testing of AS functions during the TC. The work analyzes the types of errors recorded in the database, the level of their influence on the AS, the time spent on their detection and correction. Based on the results of the error analysis, the AS functions that are subject to additional testing are selected. In this approach, the sequence of TC is considered as an iterative process of identifying possible program errors, in which new segments of tests are constructed according to the results of previous cycles. The article gives some practical results obtained in accordance with the proposed methodology and outlines new directions in the study of the effectiveness of AS testing.