Engineering an adaptive and socially- aware feedback acquisition.

Abstract

Self-adaptive software systems are characterised by their ability to monitor changes in their dynamic environment and react to these changes when needed. Adaptation is driven by these changes in the internal state of the system and its external environment. Social Adaptation is a kind of adaptation which gives users’ feedback a primary role in shaping adaptation decisions. Social Adaptation infers and employs users’ collective judgement on the alternative behaviours of a system as the main driver in tailoring adaptation decision. Users’ collective judgement is determined through individual users’ feedback collected during the lifetime of the software. Social Adaptation still lacks systematic and efficient engineering mechanisms of the acquisition process of users’ feedback. The goal of this thesis is to devise an engineering method for a systematic and adaptive acquisition of users’ feedback. Given the various contextual information which could influence how feedback should be collected from users, this thesis looks at the acquisition process itself as an adaptive process. The goal of such adaptation is to optimize the quality of obtained feedback without affecting users’ experience. In order to achieve the goal of this thesis, several empirical studies with software engineering experts and end-users have been conducted. This helped gaining insights into how the role of users’ feedback is perceived by software experts and how users behave and react to feedback acquisition. The outcomes of the empirical studies are then exploited to achieve the aim of thesis. The findings informed by these studies suggest that users’ behaviours to feedback acquisition highly varies and an adaptive feedback acquisition is highly needed to cater for differences in behaviours, improve users’ satisfaction, feedback quality and software success. To tackle this problem, the concept of Persona is employed to aid software engineers understand the various users’ behaviours and improve their ability to design feedback acquisition techniques more efficiently. The personas are developed based on a mixture of the qualitative and quantitative studies conducted throughout this thesis. In addition, this thesis proposes PAFA, a Persona-based method for a systematic design of an Adaptive Feedback Acquisition and reports on its evaluation. Finally, this thesis is also meant to contribute to the knowledge of software engineering community on developing systematic ways for feedback engineering which are hoped to lead to a better quality feedback and maintained users experience.