Designing systems that survive their authors
By jon
We’ve been quietly running a small experiment for the last quarter. The setup wasn’t dramatic and the results aren’t either, but a few people asked what we did, so this is the long version of what we did.
The thing that didn’t work
We tried, and partly succeeded, to write our reasoning down as we went. The artifacts aren’t pretty — they’re a Notion page with a lot of strikethrough — but having a paper trail of what we believed and when has saved us from re-litigating the same arguments at least twice.
Here is the part where I should give you the framework. I don’t have one. What I have is a small list of questions that, asked at the right moment, tend to prevent the worst outcomes. They’re not novel. The trick is asking them when they’re inconvenient.
The middle ground
It’s easy to write something like this and have it read as smug. I want to be honest that the version of me writing this is not the version that lived through the worst of it. That version was tired and not especially gracious.