How to select third-party library: harnessing visual insights and systematic evaluation for informed decisions

• Published in: Radìoelektronnì ì komp'ûternì sistemi (Online) Vol. 2025; no. 1; pp. 314 - 326
• Main Authors: Lysenko, Alexander, Kononenko, Igor
• Format: Journal Article
• Language: English
• Published: National Aerospace University «Kharkiv Aviation Institute 20.02.2025
• Subjects: decision-support systems; information technology; libraries migration; method; multi-criteria decision-making; radial diagrams; specification; surveys; third-party libraries; topsis
• ISSN: 1814-4225; 2663-2012
• DOI: 10.32620/reks.2025.1.20

The subject of this study is a selection approach for third-party libraries in IT projects based on the collection and analysis of expert evaluations and solving optimization problems. The aim of this study is to minimize the key challenges associated with migrating third-party libraries, including high labor intensity and significant risks that could adversely affect project success. This study sets out several objectives: 1) to create a metric specification to guide evaluation data collection through a survey, focusing on technical metrics. Also, it needs to include recommendations for surveys development; 2) formalize a two-step method that enriches the outcomes and allows for their combined use, where the first step employs radial diagrams to visualize and simplify alternative assessments, and the second step applies the TOPSIS method to address the multi-criteria decision-making problem of selecting a third-party library; 3) the practical application of the proposed approach through illustrative samples. Results. The proposed metrics specification considers the essential technical metrics and offers a unified evaluation solution. We do not limit our approach to technical metrics or numeric data. It is both possible and recommended to include human and economic metrics and to use alternatives to numeric data. The combination of multi-criteria analysis and visual methodologies simplifies the assessment and comparison of libraries, making this approach valuable for developers in practical scenarios. The conclusions of this study emphasize that its findings may lay the groundwork for developing comprehensive decision-support systems focused on third-party library selection, thereby minimizing associated risks. Future developments will include the incorporation of other metrics for library selection, as well as consideration of uncertainty and data subjectivity.