Towards Independent Software Architecture Review

Abstract

Abstract. Many software architecture evaluation methods, proposed by the research community, have a common problem of engaging the same architects to perform architecture design and evaluation. This violates the independence of quality assurance and hence may lead to biased evaluation, thereby resulting in inferior architectural design. In this paper, we analyze current approaches and issues to software architecture quality assurance. We propose seven conditions for architectural design quality assurance and discuss existing challenges towards independent software architecture design review.

Improving the Quality of Architecture Design Through Peer-Reviews and Recombination

“…An architecture review is considered as an effective way to ensure the quality of software architecture design [12,15]. However, the current architecture review methods and processes have not been widely adopted by industry due to a large number of limitations [12,14]. Historically, (formal) architecture review processes rely on time-consuming, tedious and expensive face-to-face meetings [12,14].…”

“…However, the current architecture review methods and processes have not been widely adopted by industry due to a large number of limitations [12,14]. Historically, (formal) architecture review processes rely on time-consuming, tedious and expensive face-to-face meetings [12,14]. Given the increasing trend to leveraging crowdsourcing in knowledge-intensive activities, we assert that software architecture community should explore the potential role of crowdsourcing as an alternative method in designing and evaluating software architectures and getting novice architects to gain the required knowledge, skills, and experience by soliciting the contributions from the online communities [16].…”