Home

The ogre at the end of the keyboard

7th July 2017, five minutes to read

I’m not talking about trolls on Twitter, flunkies on Facebook, or the ill-informed on the internet that have you seeing red.

I’m talking about the calm, peaceful, genuine, and amenable human at the end of a pull request or code review. They just happen to come across as the biggest, most insensitive asshole that you’ve ever had the displeasure to communicate with. And they keep doing it. All the damn time.

Gif of Ferris Bueller, Pardon my french, but you're an asshole

But you can’t avoid these jerks. Pull Requests (nee, Code Reviews) are a fundamental part of software development. Its critical that they’re productive, welcoming, and inclusive. But they’re not:

When you write and/or submit code for review, do you feel emotionally vulnerable? [poll]

— Stephanie Hurlburt (@sehurlburt) May 1, 2017

Slack recently published a great article “On empathy & Pull Requests”. It’s sentiment reminded me of the constant battle between good & evil that humans have as they give & receive feedback in the little text boxes that appear all over pull requests. Those text boxes — so innocuous, so simple — provide a direct & unshielded way to inflict harm straight to ones core, with but the tiniest barb.

Humans aren’t telepathic

When you’re reading that code from that peep millions of electrons (20 feet) away, you have the seed of an idea for a comment. You let it grow into a full formed textual representation of that idea, clear to everyone as to your intent, concern and fully conveyed context complete with subtle nuance.

Except it isn’t. It’s barely coherent, stripped of the tone in your voice, and totally lacking in those unconscious body language cues. You, in fact, sound like a insensitive & uncaring tosser, who has barely taken the time to read and understand what is right in front of you.

On top of coming across less than favorably, you have failed to actually convey the point we’re originally trying to make. The code author is hurt & angry, and you are annoyed that they can’t grasp such simple feedback. It’s so clear! Right?

Nope. Humans can’t replay exactly what the other person was thinking; we can’t read minds.

It’s clear that text isn’t the best form for conveying constructive feedback — without careful authorship, it lacks all of the non-verbal cues that make us human. But, it’s also the best we have in these situations.

This means you have to overcompensate. Not by writing more, but by writing authentically. Read it back — imagine standing next to yourself and saying the words you just typed in the most unflattering way possible. I know it wasn’t meant that way, but by golly it’ll be read that way. Then rewrite it. Use punctuation, emoji, symbols, links to make it sound like you actually meant it 😇.

A pull request is not the time to apply English 101, hoping you can make your teachers proud. Think more like those quirky back ‘n forth exchanges you read in your favourite novel.

Really, humans aren’t telepathic

You’ve written some fantastic, beautiful code. It’s the most amazing code; so simple to read & understand that anyone can follow along like a children’s book.

Except no-one else can see the matrix you are seeing. They’ve not spent 3-days understanding all the complexities of this particular component. They haven’t understood the history that led you & the code here. Or heard from the team wizard about why its fucked up the way it is, but it’s the least fucked up it could be.

To help with this, and to bring clarity to the 200 line & 4 file change, you’ve added a clear message to your commit:

Bug fixes & adding feature wibble

I mean, it’s clear from the code changes which bug fixes are present? And the Wibble.* files being added show exactly what the wibble feature is. Not only that, you’ve written copious inline code comments clarifying the why of the code, right? Thought not.

The commit & pull request descriptions are your opportunity to tell a story about about the change. It’s a chance to provide information to code archeologists when you are no longer around (or on vacation). It’s a chance to provide clarity today, right now, for this change — for this journey.

They still can’t read your mind

Even when you’ve written the 🎉best🎉 pull request message, and you’ve word-smithed your feedback to include just the right gif you still might come across as an ass. You tried your best, but some how something still got misunderstood.

Then, you, the reader, have to give them the benefit of the doubt. You have to assume that wasn’t their intent (if it was, find new coworkers & collaborators). Assume they are sweetness & light. Assume that the code is being authored with the best intentions & motivation, it has reason & logic behind it, and was intended to be that way for good reasons.

If you just can’t do that, go talk to the other person. 📞 them. IM them. Approach them from the side of you being the one at fault — not screaming down the pipe at them about being an insensitive arse. Chances are it’s just a small miscommunication.

Remember, most of all:

We’re all in this together

previous next