I like to design opinionated interfaces. That means that I am rarely just trying to solve a problem; I’m also trying to solve that problem in a specific way.
An interface is not a fact, it’s an opinion. It’s a designer’s opinion of what’s important, what the user desires, and what should be available, discernible, and instinctive.
We follow best practices for design because a lot of problems claim to be similar: users like similar patterns, do similar things, and need similar actions. But more often than not, that’s not wholly true.
It’s convenient, but it’s not precise. Users might prefer something different, but they don’t get it because it’s inconvenient to scrutinize what parts to imagine from scratch and which to keep pre-fabricated. It’s reactionary: in a world of urgency and constraints, we make the whole batch out of what’s cheap and available.
I think many of those constraints are artificial. If you want to build something timeless and beloved, you should invest as much time as is necessary to decide which parts of your interface need a strong opinion and which parts can borrow from the familiar.
Don’t assume that just because a function is mature, that its interface is also at end-stage maturity. Laptops preceded Macbooks and Blackberries preceded the iPhone. The former had a fact: you can press keys and get outputs on a GUI; you can send texts and see websites and use the calculator. The latter had an opinion: everyone should get to use a beautiful, intelligent computing device. Its interface was a direct product of that opinion, and the thing I am typing on today.