Your team’s code review process is probably an expensive illusion of quality. Here’s what the evidence says, and what to do instead.

TL;DR

  • Pull requests were designed for open source contributions from untrusted strangers. Applying them to trusted teams is a category error.
  • Peer-reviewed research shows code review’s primary value is knowledge transfer, not bug detection. Less than 15% of review comments relate to actual bugs.
  • Async PR workflows mean your code spends 86-99% of its lead time waiting. One organisation spent 130,000 hours in a single year waiting on PRs that received zero comments.
  • DORA research across 36,000+ professionals shows trunk-based development correlates with dramatically higher software delivery performance, and faster code reviews alone improve performance by 50%.
  • The alternative is T*D: Test-Driven Development (build quality in), Trunk-Based Development (integrate continuously), and Team-focused Development (review during creation, not after).
  • The transition is gradual: optimise PRs first, adopt Ship/Show/Ask, then move to pairing and trunk-based development as trust and automation mature.-
  • Kissaki@programming.devEnglish
    4·
    13 days ago

    I would rather not have my 20-60 work commits, possibly shifting solution design twice or three times, in the main project history, spanning days and sometimes weeks, invalidating earlier commits, and a hassle if not impossible to reasonably and efficiently document changes between them. And all those changes intertwined with my other work and my colleagues’ changes.

    The key insight from Charity Majors is that speed and safety are not trade-offs – speed IS safety. “Ship a single changeset by a single dev at a time, making it easy to isolate the owner of any problem, preventing the blast radius from expanding, and making it easy to fix while the intended effects of the code are fresh in their mind.”

    I really don’t get how “develop in feature-flagged trunk” is supposed to come together with “single dev single changeset”. Work evolves. Design evolves. Surely, multiple/many “single changesets” must come together to form new features and changes? This claim seems to assume issues are found before it’s being changed again, and in a timely manner, even when it is behind a feature flag?

    Were pull requests really “created for FOSS”? I feel like, maybe assume, pre-merge reviews have been a thing before that [and unrelated to PRs].

    I feel like the article could have been structured a lot better, and made its point without a “it’s like this” “but maybe not” digression. Feels like inflammatory bait into relativation. I guess I’ll refrain from raising my other points and disagreements that I noted because by the end of the article, it wasn’t really claiming it broadly anymore anyway.

    Outside of the irritation, I found the arguments interesting. I’ll keep it in mind, but I don’t see my team implementing it like that or to a higher degree. For various reasons.

    Despite the negative vote sum, I think it’s an interesting alternative perspective. Even if I found it irritating to read, and even if people hate it, it’s a good discussion in the comments. About how we work.