How can feature usage be tracked across product variants? Implicit Feedback in Software Product Lines

Abstract

Implicit feedback involves gathering information about software usage to grasp how and when the software is utilized. This study investigates the integration of implicit feedback mechanisms into Software Product Line Engineering (SPLE). While common in product-based development for identifying bugs, usability issues, and informing requirement prioritization, its adoption in SPLs has been limited due to the unique challenges posed by SPLE, such as the emphasis on Domain Engineering over Application Engineering, and the need for systematic reuse of shared assets. We propose a novel approach to incorporate feedback practices into Domain Engineering, thereby shifting the focus from individual product variants to the SPL platform, and specifically moving from product-based feedback to feature-based feedback. Based on a case study, we suggest that product derivation includes a second step that injects the trackers at the time of derivation, using a Feedback Model that complements the Configuration Model for feedback analysis.To test this approach, we introduce FEACKER, an extension to pure::variants as the variability manager. FEACKER injects trackers when the product variant is derived. The findings are validated through a Technology Acceptance Model (TAM) evaluation and a focus group discussion, providing insights into the feasibility, acceptance, and potential impact of platform-based feedback in SPLE. The results indicate agreement on the benefits of conducting feedback analysis at the platform level and the perception that FEACKER seamlessly extends the capabilities of pure::variants.