Compromise, design and the literal edges
Let's say you wanted to improve the katana, the legendary fighting sword.
You could ask your team to come up with a sword that's lighter, sharper and more durable.
Built into that charge is the requirement to compromise. And just about everyone who has come before you has tried to come up with the same sort of compromise, and your chances of a breakthrough are slim indeed.
Compromise gives us an out, because, with multiple goals, it's easy to play it safe.
But what if you picked just one?
What if you sought to make the sharpest katana ever? Or merely the most durable one? By optimizing for just one attribute, you've eliminated most of the compromise from the design discussion. As a result, you're far more likely to encounter something extraordinary. It might not be practical, but there's plenty of time to compromise later.
It's almost always easier to roll something back a little than it is to push it forward.