Non-Blocking Continuous Code Reviews, a Case Study - Thierry de Pauw - NDC Oslo 2026

Value of the Information & Strength of the Argument

The value of the talk lies in its concrete, real-world implementation of a non-blocking review process, offering a pragmatic alternative to the prevalent pull-request model. The argumentation is strong, systematically deconstructing the costs of blocking reviews (transaction costs, work-in-progress, delayed feedback) and showing how the case study avoided these. The speaker supports his points with references to lean manufacturing principles (transaction costs, Little’s law) and mentions empirical research (gender bias paper, Accelerate), which bolsters credibility. However, the argument is primarily driven by personal experience and might lack generalizability, and some claims (e.g., ‘code reviews don’t catch bugs’) are presented categorically without nuance. The logical flow from problem identification to solution and benefits is clear and persuasive, though occasional tangents about organizational culture could distract from the core technical message.

Scientific Rigor, Source Quality, Title Accuracy

The speaker demonstrates scientific rigor by citing relevant research (such as the 2016 gender bias in GitHub pull requests study and the book Accelerate) and using established concepts (transaction costs, Little’s law). However, these references are mentioned anecdotally without detailed citations or URLs, limiting the ability to verify. The presentation’s title accurately reflects the content: it is indeed a case study of non-blocking continuous code reviews. The speaker also acknowledges the limitation of his own experience, noting that he initially thought the approach was not notable until others showed interest. Overall, the talk is grounded in practical experience and some theory, but the lack of explicit external references in the transcript or description reduces the strength of source verification.

The title accurately describes the talk's focus on non-blocking continuous code reviews, illustrated through a specific team case study.

7/10

The talk presents a coherent case study with some references to research (gender bias paper, Accelerate), but it is primarily based on personal experience and practitioner opinion. The arguments are well-structured and supported by concrete examples, though some claims lack external verification.

Markers derived by PSI from the transcript: the creator did not define chapters.

• 02:00Introduction to the talk and speaker background.
• 08:00Discussion on the purposes of code reviews: mentoring, knowledge sharing, maintainability.
• 17:00Explanation of why process gates reduce quality and create trust issues.
• 26:00Case study setup: 2012 novice team, no version control, introduction of trunk-based development.
• 37:00Key idea 1: establishing a continuous review process with a Kanban 'to review' column.
• 43:00Key idea 2: decoupling code review from release decisions, allowing unreviewed code to go to production.
• 51:00Benefits analysis using transaction costs, reduction of work in progress, and improved feedback.
• 58:00Conclusion and emphasis on team culture and potential challenges.

• NDC Conferences — Conference organizer; the talk was presented at NDC Oslo.
• NDC Oslo — Conference venue and event information.