Socio-Technical Smell Dynamics in Code Samples: A Multivocal Review on Emergence, Evolution, and Co-Occurrence

Abstract

Code samples play a pivotal role in open-source ecosystems (OSSECO), serving as lightweight artifacts that support knowledge transfer, onboarding, and framework adoption. Despite their instructional relevance, these samples are often governed informally, with minimal review and unclear ownership, which increases their exposure to socio-technical degradation. In this context, the co-occurrence and longitudinal interplay of code smells (e.g., large classes, poor modularity) and community smells (e.g., lone contributors, fragmented communication) become particularly critical. While each type of smell has been studied in isolation, little is known about how community-level dysfunctions anticipate or exacerbate technical anomalies in code samples over time. This study investigates how code and community smells emerge, co-occur, and evolve within code samples maintained in OSSECOs. A Multivocal Literature Review protocol was applied, encompassing 30 peer-reviewed papers and 17 practitioner-oriented sources (2013-2024). Thematic synthesis was conducted to identify recurring socio-technical patterns related to smell dynamics. Nine patterns were identified, showing that community smells often precede or reinforce technical degradation in code samples. Symptoms such as "radio silence" and centralized ownership were frequently associated with persistent structural anomalies. Additionally, limited onboarding, the absence of continuous refactoring, and informal collaboration emerged as recurring conditions for smell accumulation. Conclusion: In OSSECOs, particularly within code samples, community-level dysfunctions not only correlate with but often signal maintainability decay. These findings underscore the need for socio-technical quality indicators and lightweight governance mechanisms tailored to shared instructional artifacts.