QUALIMETRY OF SOFTWARE REQUIREMENTS AT THE INITIAL STAGES OF THEIR LIFE CYCLE

Abstract

This article addresses the scientific and practical problem of quantitative assessment of software requirements (SR) at the initial stages of their life cycle. Traditional approaches lack sufficient objectivity in the requirements analysis process, as requirements are often qualitative and subjective, which creates high risks of inefficient resource allocation and planning errors. The proposed qualimetric approach is based on a scaling methodology designed to convert non- numerical characteristics of requirements into precise quantitative values, a necessary condition for their mathematical analysis and comparison. The central element of this work is the developed method for creating requirements scales, which formalizes the expert evaluation process and allows for assigning a numerical rank to qualitative criteria. Specifically, a concrete scale for the “Necessity” characteristic of a requirement was created and justified. This scale allows for standardized assessment of the requirement’s criticality, clearly distinguishing between mandatory, important, and desirable functional/non-functional aspects of the SR. This forms the basis for effective prioritization. For the integral assessment of complex, multi-factor requirement properties, the use of the weighted mathematical expectation apparatus has been implemented. This mechanism enables the calculation of complex characteristics by taking into account not only the acquired scores but also the weight (significance coefficient) of each individual criterion, thereby ensuring a realistic and comprehensive assessment of the overall “value” of the requirement for the project. The practical significance of the work lies in enhancing the transparency and justification of the requirements management process, allowing development teams to make informed decisions regarding the scope of work and iteration planning even before the active programming phase begins. The scientific novelty lies in the adaptation and systematization of the qualimetric apparatus for the specific domain of requirements engineering at the early stage of the life cycle. Perspectives for further research include: the development of new scales for additional key characteristics (e.g., “implementation complexity”, “risk level”), empirical validation of the proposed methodology on various real-world projects to confirm its universality, and the automation of the assessment and calculation process through the creation of software tools integrated with requirements management systems. Also promising is the investigation of methods for dynamic priority recalculation under changing conditions, based on fuzzy logic or machine learning.