💾 Archived View for r.bdr.sh › gemlog › 1724329272429 › pull-requests-are-overrated.gmi captured on 2024-08-25 at 00:00:35. Gemini links have been rewritten to link to archived content

View Raw

More Information

-=-=-=-=-=-=-

Pull Requests Are Overrated

I don't think pull requests are all that useful in product teams, and often they are signs of mistrust, or hide issues earlier in the process. Sure, a lot can be automated to make pull requests simpler, but even things that aren't easily automated can be done better earlier on.

If you're working in a team, you shouldn't be designing alone: you should be discussing and validating with your peers before a single line of production code is written. If the first time someone is seeing your concept is during the pull request, then it's already too late:

• It will take longer to understand. Trying to understand the concepts and structure of a program by reading the code takes more effort than using prose or diagrams, this makes the review process more time consuming.
• Changes will be more costly. It's a lot easier to change a design on paper than it is to do it on code, this means that any big change will require much more work.
• The pressure of deadlines will be higher. If you're challenging the design of a solution later in the process, it's likelier to not be taken in "so we can release on time". This leads to a lot of backlog-filler improvement projects and unaddressed design issues.
• Anchoring will prevent exploring alternatives. Combining the three above, you're likelier to make tradeoffs that focus on adjacent solutions instead of truly exploring better ways to build.

Most teams want to be fast at solving problems, and one of the best ways to do that is to slow down where it matters: Be thoughtful and intentional with your design, understand your problem well, go broad with ideas early in the process when it's easy and cheap to do it. Stop writing code until you know exactly what code you'll write. In my experience, this is the most valuable change you can do, and it benefits engineers at every career level.

Guardrails prevent us from making mistakes, but they also allow bad habits to form. What would happen if your team decides to stop using Pull Requests today? How can you address those issues without them?