Essay · 4 min read
Notes on doing one thing well
Most tools try to do everything. The ones I keep coming back to do one thing, and they do it the same way every time.
A good pencil doesn't ask what you want to draw. It gets out of the way, and the line comes out where your hand meant it to.
Software rarely works like that. Every release adds a setting, every setting adds a question, and after a while the tool is mostly questions.
The alternative isn't fewer features for their own sake. It's deciding what the thing is for, and letting that answer most of the questions before anyone has to ask them.
When a tool knows what it's for, the details start to line up. The button is where your thumb already is. The default is the choice you would have made.
None of this shows up in a feature list. It shows up as the absence of friction, which is hard to sell and easy to feel.
It also shows up in what doesn't happen. No dialog asking if you're sure, because the action can be undone. No empty state, because the first screen already has something in it.
Restraint like this is slow work. Adding is quick and feels like progress; taking away means knowing the thing well enough to see what it doesn't need.
That's why the simplest tools are so often the oldest. They've had time to lose everything that wasn't carrying weight.
So the next time something feels simple, look closer. Somebody probably spent a long time deciding what to leave out.
Back to top
NewFills its ring as you read, and when you press it the arrow lifts off while the page glides back up.
Navigation