Janis Bicevskis

SMOTL—A System to Construct Samples for Data Processing Program Debugging

The possibility of automatic construction of a complete set of program tests is considered. A test set system is said to be complete if every feasible program branch (segment) is executed by it. The complete test set construction algorithm for commercially oriented data processing programs is outlined, and the results of its functioning on real programs are analyzed.

Automatic Construction of Test Sets: Theoretical Approach

We consider the problem of automatic construction of complete test set (CTS) from program text. The completeness criterion adopted is C1, i.e., it is necessary to execute all feasible branches of program at least once on the tests of CTS. A simple programming language is introduced with the property that the values used in conditional statements are not arithmetically deformed. For this language the CTS problem is proved to be algorithmically solvable and CTS construction algorithm is obtained. Some generalizations of this language containing counters, stacks or arrays are considered where the CTS problem remains solvable. In conclusion the applications of the obtained results to CTS construction for real time systems are considered.

Requirements Model of Sociotechnical Systems Simulator Architecture

To avoid any voluntary decisions it is desirable to examine the possible consequences that can be done by simulation. Unfortunately there is no unified methodology and technology for simulators designing. Above mentioned problem when development of each simulator has to be started from the scratches makes elaboration expensive and cumbersome. The article dealt with the way to joint understanding of simulators designing concept.

Asynchronous Runtime Verification of Business Processes

The authors propose a runtime verification mechanism for business processes. This mechanism allows verifying the correctness of business process execution and it runs in parallel with the base processes affecting them insignificantly. The authors have identified the case where the use of business process runtime verification is helpful and applicable. The verification mechanism monitors the business process execution and verifies compliance with the base process description. The verification mechanism prototype was developed and tested in real business processes, as well as limits of runtime verification overhead were evaluated.

Models of Event Driven Systems

This paper provides the business process modeling approach based on usage of Domain Specific Languages (DSL). The proposed approach allows us to create executable information systems’ models and extends the concept of Event Driven Architecture (EDA) with the business process execution description. It lets us apply principles of the Model Driven Development (MDD) in order to create the information system which complies with the model. The proposed approach provides a set of advantages in information systems development, use and maintenance: bridges the gap between business and IT, an exact specification, which is easily to implement into information system, up-to-date documentation etc. The practical experience proves the viability of the proposed approach.

Smart technologies for improved software maintenance

Steadily increasing complexity of software systems makes them difficult to configure and use without special IT knowledge. One of the solutions is to improve software systems making them “smarter”, i.e. to supplement software systems with features of self-management, at least partially. This paper describes several software components known as smart technologies, which facilitate software use and maintenance. As to date smart technologies incorporate version updating, execution environment testing, self-testing, runtime verification and business process execution. The proposed approach has been successfully applied in several software projects.

Domain-specific characteristics of data quality

The research discusses the issue how to describe data quality and what should be taken into account when developing an universal data quality management solution. The proposed approach is to create quality specifications for each kind of data objects and to make them executable. The specification can be executed step-by-step according to business process descriptions, ensuring the gradual accumulation of data in the database and data quality checking according to the specific use case. The described approach can be applied to check the completeness, accuracy, timeliness and consistency of accumulated data.

Models of Data Quality

The research proposes a new approach to data quality management presenting three groups of DSL (Domain Specific Language). The first language group uses concept of data object in order to describe data to be analysed, the second group describes the requirements on data quality, and the third group describes data quality management process. The proposed approach deals with development of executable quality specifications for each kind of data objects. The specification can be executed step-by-step according to business process descriptions, ensuring the gradual accumulation of data in the database and data quality verification according to the specific use case.

Self-management of Information Systems

The paper discusses self-management features that are intended to support the usage and maintenance processes in the information system life. Instead of a universal solutions that are evolved by many researchers in the autonomic computing field, this approach, called smart technologies, anticipates self-management features by including autonomic components into information systems directly. The approach is practically applied in several information systems, and the gained results show that the implementation of self-management features requires relatively modest resources. Thereby the approach is suitable even for smaller projects and companies.

Transparent Model for State Management

To ensure optimum governance of the state (regions, sectors, specific activities) all management system units (tasks, functions and services, normative environment, institutional framework, budget funding) have to operate in strong cooperation and consistency, which should be defined by corresponding documents. Unfortunately this logical relevance does not always exist. Documents and information are weakly connected, because of the huge complexity of units that objectively hinders implementation of the strong linkage and seriously weakens management quality. The situation can be radically improved by usage of ontological methodology for development of the back-office of the management system—strict structuring of objects and their decomposition in elementary logical units to form different hierarchies of information. Each hierarchy displays some aspect of information. Connection of related units of the information from different documents (hierarchies) creates a full graph of relevant information; it enables processing and usage of management information in static and dynamic regimes—definition and concretization of normative acts, institutional structure, funding—passage and execution of tasks, etc. Analysis of information from different aspects, checking its consistency, and making appropriate decisions on this base is becoming possible, thus minimizing threats to well-functioning and development of the country. Two case studies illustrate the current usage of the proposed principles. Complexity of the informative model requires appropriate level of IT support. Requirements as well as short descriptions of developed IT tools are included; they are oriented on usage by non-IT specialists (civil servants). Projections of further activities for improvement features of the model and consummation of IT tools are mentioned.