27 February 2025
14:00 Master's Defense Room 85 of IC2
Topic
On the adoption of Continuous Practices: Challenges and future solutions
Student
André Luiz do Canto Portela
Advisor / Teacher
Breno Bernard Nicolau de France
Brief summary
Context: Continuous Software Engineering and DevOps rely heavily on Continuous Practices (CPs) such as Continuous Integration, Continuous Delivery, and Continuous Deployment to improve software reliability, business adaptability, and reduce time to market. Although such PCs can significantly improve software development processes, organizations still face challenges in their adoption. These challenges can be categorized as Social Challenges or Technical Challenges (DTs). Objective: Organizations face several DTs when adopting PCs, and it is essential to mitigate them. This dissertation focuses on identifying DTs and their detection strategies, considering both Academic and Industry perspectives. Methods: We conducted a Rapid Review (RR) focusing on practical problems faced by professionals to identify DTs. We then conducted a Case Study (CS) in collaboration with a software engineering consulting firm and developed a conceptual framework for detecting DTs. RR focused on reviewing relevant scientific literature, in collaboration with professionals, applying a quality assessment of the selected studies and a narrative summary to identify and categorize DTs and their respective detection strategies. EC provided real-world insights through interviews, meetings, and expert feedback. The interviews were analyzed using Thematic Analysis (TA), and the DTs and categories identified in the TA were compared and integrated into the RR catalog. Additionally, we held biweekly meetings with experts over the course of a year to validate the identified DTs and develop a static analysis-based DT detection framework. Results: RR revealed 121 DTs organized into nine categories and identified detection strategies for 100 DTs, providing a basis for future work. To bridge the gap between academic findings and practitioner needs, the results were summarized in an Evidence Briefing, a document designed to inform software practitioners. The EC generated additional contributions. 31 challenges were identified, including 25 DTs and 6 Social Challenges. EC has expanded the RR challenge catalog with 26 new challenges and four additional categories. These findings highlight the complexity of PC adoption in real-world contexts and emphasize the need to address technical and social barriers. The end result of this research is a framework for detecting DTs using static analysis in software repositories, which enables organizations to address DTs before they impede PC adoption. Conclusion: We identified differences in the perspectives of academia and industry, since categories such as Infrastructure and Technology can represent significant DTs. While some DTs have a low likelihood of completely preventing CPs, others can significantly impact their adoption and the organization as a whole. This research expanded the catalog of known DTs, providing a concrete basis for future studies. The proposed framework has the potential to enable faster and more reliable detection of DTs; however, its practical effectiveness still needs to be evaluated.
Examination Board
Headlines:
| Breno Bernard Nicolau de França | IC / UNICAMP |
| Paulo Sérgio Medeiros dos Santos | UNIRIO |
| Bruno Barbieri de Pontes Cafeo | IC / UNICAMP |
Substitutes:
| Leonardo Montecchi | IC / UNICAMP |
| Carla Silva Rocha Aguiar | UnB |