Dmitrii Iastrebov’s scientific contributions

Projects of modern software systems comprise a wide number of various design artifacts. There is an extensive set of approaches and tools for creating design artifacts. It is very important to use good practices to improve the quality of design artifacts and a development process of creating software systems. Modern software project repositories contain many software systems. However, there are no effective methods for information retrieval and analyzing existing projects to gain access to best practices.Thus, it is necessary to develop a model of a software system project, considering various approaches and tools for a design artifacts creation and a development process organization. Such a model will make it possible to search for software projects, considering their domain features and properties of design artifacts and development process. Various data mining and systems analysis methods can apply to a found set of projects to extract a set of best practices.We described the approach to building the information retrieval module for the intelligent design repository in this paper. Also, we presented the model of a software system project. We considered the algorithm for indexing a software system project to build the information retrieval module index. There is a measure of a distance between a search query and an index element of the information retrieval module. We proposed the algorithm for calculating the relevance of an index element to a search query. The article also provides examples of the work of the information retrieval module for the intelligent design repository.KeywordsInformation retrievalData miningSoftware repository

A modern large software organization has significant experience in the design and development of software systems. That experience was obtained in the process of working on previous (completed) projects. Software project repositories contain the experience and knowledge of a large number of highly qualified specialists for many years of their work. Software project repositories contain heterogeneous unstructured data. It becomes difficult to find the necessary information in the process of new project development. So the following problems exist: the experience of previous projects remains unclaimed, the time of the development process increases, and the quality of the final product decreases. So it is fundamentally important to use the experience of the previous projects at the initial stages of new projects development. The usage of intelligent methods and algorithms for software project repositories analyzing is the solution to this problem. This article proposes an approach to the data mining of software project repositories based on features of the domain and the development process dynamic. This approach allows to reduce the search space and speed up the process of the required knowledge searching to the new project development.