Yeah, step 1 is wrong too. The article goes into that.
You can't design an interface based on a partial feature-set, you need full interaction modes specified beforehand. You can't finish the details before you implement it, or even before you test. You can't have a committee of "everyones" to like it before you test.
Combining the steps isn't something "you can do", it's the one way that works.
You can't design an interface based on a partial feature-set, you need full interaction modes specified beforehand. You can't finish the details before you implement it, or even before you test. You can't have a committee of "everyones" to like it before you test.
Combining the steps isn't something "you can do", it's the one way that works.