11.1. The main factors determining the quality of complex software

General understanding of the quality of PS international standard ISO 9126: 1-4: 2002 it is recommended to describe the three interacting and interdependent metrics of quality characteristics, reflecting:

• internal quality, which is manifested in the development process and other intermediate stages of the PS life cycle;

• external quality specified by customer requirements in the specifications and reflected by the characteristics of the final product;

• quality when used during normal operation and the effectiveness of achieving user needs, taking into account the cost of resources.

Internal metrics in accordance with the standards can be applied during the design and programming of the PS components, such as a specification or source program text. When developing PS, the intermediate components should be evaluated using internal metrics that reflect the functional and structural properties of the programs. The main purpose of using internal metrics is to ensure that developers get the required external quality. It is recommended to use internal metrics that have the strongest links to priority external metrics, so that they can help in predicting their achievable values. Internal metrics enable developers, testers and customers, starting with system design, forecasting the quality of the life cycle of the programs and the issues of technological quality assurance before the PS becomes a ready-to-use product. Measurements of internal metrics use properties, categories, numbers, or characteristics of PS elements, which, for example, are found in the source program text procedures, in the control flow graph, in the data flow, and in descriptions of changes in memory states.

External metrics use PS measures that reflect the behavior of the system of which they are a part, through testing, operating, and observing executable programs or operating the system. Before purchasing or using a PS, it should be evaluated using metrics based on the realization of business and professional goals related to the use of a software product in a specific organizational and technical environment. External metrics provide customers, users and developers with the ability to track and analyze the quality of PS during testing or trial operation. Suitable external metrics are specified to obtain numerical values ​​or categories and properties of internal quality characteristics, so that they can be used to verify that intermediate products in the development process meet internal quality specifications.

Quality metrics in use reflect the extent to which a product meets the needs of specific users in achieving specified goals. This metric is not reflected in the number of six basic characteristics of the PS, which are regulated by the ISO 9126-1 standard due to its generality, however, it is recommended for an integrated assessment of the results of operation and the use of the program complexes in ISO 9126-4. The relationship of quality in use with other PS characteristics depends on the tasks and functions of their consumers (see Lecture 6).

Standard ISO 9126: 1-4 - it is advisable to use as a basis for formal regulation of the quality characteristics in the life cycle of software projects. The model of the quality characteristics of the PS and components consists of six groups of basic indicators, each of which is detailed by several regulatory sub-characteristics.

Functionality is detailed:

• suitability for intended use;

• correctness (correctness, accuracy) of the implementation of requirements;

• ability to interact with components and environment;

• security - safe operation. Reliability is characterized by:

• the level of completion - the absence of defects and errors;

• stability in the presence of defects and errors;

• recoverability after the manifestation of defects;

• availability - the willingness to implement the required functions. Efficiency is recommended to reflect:

• temporary effectiveness of the implementation of a complex of programs;

• usability of computing resources. Applicability (practicality) is proposed to describe:

• clarity of functions and documentation;

• ease of use of a set of programs;

• study of the processes of functioning and application. Accompanying submitted:

• analyzability - convenience for analyzing the proposals of modifications;

• the variability of components and complex programs;

• testability changes when accompanied. Mobility (portability) is proposed to reflect:

• adaptability to environmental changes;

• ease of installation - installation after the transfer;

- substitutability of the components during the program complex adjustments.

Characteristics and subcharacteristics in the standard are defined briefly, without comments and detailed recommendations for their application to specific systems and PS projects. The presentation is conceptual in nature and does not contain recommendations for the selection and streamlining of priorities, as well as the necessary minimum of criteria depending on the characteristics of the object, development environment, maintenance and application.

To select the quality characteristics of the PS and reliably compare them with the requirements, as well as to compare their values ​​between different

These software products require assessments, measurements, and use of certain measures and scales. Standards it is recommended that the measurement of each quality characteristic of the PS (subcharacteristics or its attribute) be provided with an accuracy and certainty sufficient for comparison with the requirements of technical specifications and specifications, and that the measurements are objective and reproducible. The norms of permissible measurement errors caused by tools and / or errors of a human expert should be provided for. In order for the measurements to be objective, a procedure must be documented and agreed upon for assigning a numerical value, property, or category to each attribute of the software product. The characteristics, subcharacteristics and quality attributes of the PS from the position of possibility and accuracy of their measurement can be divided into three levels of detail indicators, the features of which should be clarified when choosing them:

• categorical-descriptive, reflecting a set of properties and general characteristics of an object — its functions, categories of responsibility, security and importance, which can be represented by a nominal scale of categories of properties;

• quantitative - represented by a set of ordered, numerical points reflecting continuous or discrete patterns and described by an interval or relative scale that can be objectively measured and compared numerically with the requirements;

• Qualitative - containing several ordered or separate properties - categories that are characterized by an ordinal or point scale of a set of categories (yes, no, good, bad) are established, selected and evaluated largely subjectively and expertly.

The first level includes the quality indicators, which are characterized by the greatest variety of values ​​- the properties of programs and data sets and cover the whole range of classes, purposes and functions of modern PS. These properties can be compared only within the same type of PS and it is difficult to order according to the principle of preference. Among the standardized quality indicators, this group primarily includes functional fitness, which is the dominant characteristic of any PS. The nomenclature and values ​​of all the rest quality indicators are directly determined by the required functions of the software and, to varying degrees, affect the performance of these functions.

Functional suitability is the most responsible, objectively difficult to formalize and evaluate the characteristics of program complexes in the project. This characteristic is related to what basic and additional functions and tasks the software should solve to meet the needs of users, while other design characteristics are mainly related to how and under what conditions the specified functions can be performed with the required quality. Subcharacteristics and attributes of functional suitability can be characterized mainly by features, categories and qualitative description of functions for which it is often difficult to determine numerical measures and scales.

The second level of quality indicators are fairly reliably and objectively measured numerical characteristics of PS. The values ​​of these design characteristics usually have the greatest effect on the functional suitability in the use of PS. Therefore, the choice and justification of their required values ​​should be carried out most accurately and reliably already in the design of PS. Their subcharacteristics can be described by ordered scales of objectively measured values, the required numerical values ​​of which can be set and selected by customers or users of the software. Such characteristics are the reliability and efficiency of the program complexes. These values ​​can be selected and recorded in a technical specification or specification of requirements and be accompanied by a method of objective, numerical measurements during qualification tests for comparison with requirements. The duration of solving basic problems, the capacity for the number of their solutions for a certain time interval, the duration of waiting for results (response) and some other characteristics of the dynamics of the PS operation can be selected and set quantitatively in the specifications of customer requirements.

The third level of standardized PS quality indicators is difficult to fully describe with measured quantitative values, and some of their subcharacteristics and attributes have descriptive, qualityny view Depending on the functional purpose of the PS, in agreement with the customer, it is possible to determine expertly the degree of necessity (priority) of these properties and the scoring values ​​of the level of realization of their attributes in the life cycle of a specific PS.

The problem is to identify the factors on which they depend, to create methods and means of reducing their influence on the functional suitability of PS, as well as to efficiently allocate limited resources to ensure the required quality of functioning of a complex of programs that is equally strong for all real negative impacts. The complex, coordinated application of these methods and means in the process of creating, developing and applying PS makes it possible to exclude manifestations of a number of negative factors or significantly weaken their influence. Thus, the level of achieved quality of PS operation can be predictable and manageable, directly dependent on the resources allocated for its achievement, and most importantly, on the quality system and technology efficiency used at all stages of the PS life cycle.