Thoughts on Pull Requests - Description

Having been a developer now for quite a bit, I’ve developed a sense of what is important to me. It is not about the technology I use, it is not about the part of development I do. It is about systems and processes that not only set myself up for success, but also my team and future members. One crucial part that I care about is the pull request.

The pull request is a mechanism for other folks to see your code, make suggestions, and extend your codebase/product/etc. In some companies, your pull request is connected to a specific ticket for project tracking purposes. This allows folks to trace how code evolved either to fix a product issue, extend the system, or even for ways to refactor the codebase. Some folks, however, don’t think they matter that much and will just be light on their feedback or simply put little to no effort in describing their change beyond the code itself.

The pull request to me is incredibly important. The pull request allows us an opportunity to show the intention. We are able to see what a set of changes will do to our codebase and how it is written even in the description is crucial. If we’re asking for folks to review our code, it is important we give folks as much context as they need to understand the set of changes. As the person opening it we need to make sure we put our perspective away and write it from someone coming in who might not know the full picture and how we can give them enough that they can feel confident the change will perform the way it should.

What’s in the Pull Request Description?

To facilitate that, we should make sure to always include a helpful description. There are a few key elements that matter to me that I try to include in every pull request:

Summary and Before/Afters

A summary of the intention. This can be brief just to give the highlights of what the work will do. Generally for user interface work, I think it is also important to highlight any visual changes in this section as a before/after or even just the way the interactions should work as a small preview for folks so they don’t have to pull the entire thing down and make it run.

Technical Details

From there I generally have a section around any “technical detail” where I can dive deeper into the implementation or anything in particular to articulate the complexities. This is generally where I might explain a bit about the system or any necessary backend calls that we’re listening to. Or the complexity of refactoring the code.

Testing Instructions

Most importantly to me, there is the section for how to actually go about testing the code change manually. In my jobs, we have developers test our code and so writing down the interactions can be really useful — especially if they need steps for how to boot the application in a certain way. This also is used for our quality assurance processes and so I can send it to our QA engineers to manually test it or communicate it our automation teams.


And the last bit I usually do is any articles or tickets I should reference for the work. I am generally unique in this way but I care deeply about sharing resources for people if I am implementing functionality that is weird or new to folks. Or if I’m using any specific APIs from SDKs we either control or don’t control. This is CRUCIAL to me because it allows us make sure future developers understand where the problem came from or what tool we were operating with. Adding references allow us to also use the pull request as a means of education and expanding wisdom. Back when I worked in a Design Systems team, we used this section as a way to also highlight articles around accessibility or other design systems we referenced to understand how we wanted to architect the developer experience.

These are some of my thoughts around the description. This is the very first thing folks see. It is also the thing that ends up potentially getting added into your squash commits or merge commits. So having all this historical information at the ready for people can save them lots of time!