This is the door for an office I do some work at. See the little signs near the handles that say “Push”? The signs weren’t there to begin with but so many people were trying to open it by pulling, that the folks installed the signs.
But we didn’t need a sign. We needed a door that didn’t need a sign.
Look around. In *almost every situation you see a sign, there’s a design failure. The sign doesn’t solve the problem. It treats the symptoms of the problem.
l’m not saying never put up a sign. Let’s be pragmatic. Sometimes, when the problem is temporary or the cost to fix is unjustifiable, putting up a sign is the best choice.
But at some point in the future, you will find yourself in a situation where you are tempted to solve a problem by putting up a sign. And that’s when I ask you to think: is there an underlying issue that I should be fixing instead?
(By the way, software designers, this goes for you too. Your “signs” might be anything from an alert box or a tooltip message, to comments in your code)
*The main exception I can think of are road signs but even then, a fair number of road signs could be eliminated with better designed intersections and exchanges. We likely don’t even know all the places where we can eliminate signs. 30 years ago, nobody would have thought you could hop on a new computer and learn to use it without reading any instructions, but that’s exactly what people do with iPads today. Great design explains itself.