Analysis of non-functional requirements for smart contract recommender system
DOI:
https://doi.org/10.31449/upinf.222Keywords:
Ethereum virtualni stroj, nefunkcionalne zahteve, orodja umetne inteligence, trilema tehnologije veriženja blokovAbstract
Non-functional requirements (NFRs) are essential criteria of software quality that affect the performance, usability, security, and reliability of software. The blockchain, as a novel technology expands the set of common NFRs with additional attributes, such as transparency, immutability, and decentralization. However, blockchain integration usually focuses on functional requirements (FRs), while NFRs are not sufficiently addressed. Therefore, NFRs are usually analysed with dedicated APIs, crawlers, and other tools that require ongoing maintenance and manual intervention. This paper proposes an AI-based analysis of NFRs for blockchain solutions, where AI tools such as Copilot, ChatGPT, and others are leveraged. We use these tools to generate, evaluate, and optimize NFRs for blockchain solutions in a more comprehensive way. The results of our solution are presented on a recommender system use case, including explanation on how to improve the sustainability of the proposed solution.