Design is mostly about synthesizing many different goals and priorities into a coherent approach. It's pretty easy to take one goal or activity and design a flow for it, but making it work with all the other objectives of a project will take a lot of compromise. To make a simple and usable UI, you will need to design chunks that work well enough for a half-dozen different activities.
It's similar to architectural design.
It would be relatively easy for somebody who is not an architect to design a room built for one purpose only, say reading. It would have rows upon rows of ergonomic chairs with cupholders and book or ereader holders. This room would hold many people who could all read at once. It would be useless for anything else. Also, people tend to become uncomfortable in such a mechanistic and single-minded environment.
Most would rather read in a coffee shop, or a living room, rooms designed with compromises and artistic sense to support many, many different types of activities. I'm not going to get into specifics, but cubicle farms have the same problem.
The second most important part is painstaking logic. Once you have an overarching concept that unifies all the goals of your UI, you need to step through all of the most important activities (or "flows"), trying to determine whether it is the easiest possible way to complete the action, and if you can make it easier without breaking the overarching design. Sometimes you can't and that's OK. But if you can make it easier, you should.
Then there's visual design- This is tough because it is a completely separate discipline from UX, yet it is intimately bound to it. The main thing to remember is visual weight. It's bad if everything has the same weight. A good rule of thumb is the 1-2-3 presentation. There should be some elements that are primary, some that are secondary, some that are tertiary. It goes without saying that these weights correspond to the importance of certain actions in the interface. Play mind tricks on your user to steer them towards desirable behaviors.
If you can take something out without too many negative consequences, do it, because that will give you room later to make a completely different flow easier. Or leave that room empty to make your interface more simple and beautiful.
As for "scientific theories of design": Most of these things deal with visual weight and local presentation and have been common sense to artists for millenia, and are about as useful as A/B testing. That is, they are very useful towards the end to make sure that all of the separate elements in your design have reached local maxima. Maybe it's a bit like photography- if the photographer can give you scientific reasons for his white balance and exposure etc., good for him, but they say nothing about the subject matter or even the composition of the shot.
The ultimate criteria is the response of the audience, and no amount of tests or theories will ever touch that.
It's similar to architectural design.
It would be relatively easy for somebody who is not an architect to design a room built for one purpose only, say reading. It would have rows upon rows of ergonomic chairs with cupholders and book or ereader holders. This room would hold many people who could all read at once. It would be useless for anything else. Also, people tend to become uncomfortable in such a mechanistic and single-minded environment.
Most would rather read in a coffee shop, or a living room, rooms designed with compromises and artistic sense to support many, many different types of activities. I'm not going to get into specifics, but cubicle farms have the same problem.
The second most important part is painstaking logic. Once you have an overarching concept that unifies all the goals of your UI, you need to step through all of the most important activities (or "flows"), trying to determine whether it is the easiest possible way to complete the action, and if you can make it easier without breaking the overarching design. Sometimes you can't and that's OK. But if you can make it easier, you should.
Then there's visual design- This is tough because it is a completely separate discipline from UX, yet it is intimately bound to it. The main thing to remember is visual weight. It's bad if everything has the same weight. A good rule of thumb is the 1-2-3 presentation. There should be some elements that are primary, some that are secondary, some that are tertiary. It goes without saying that these weights correspond to the importance of certain actions in the interface. Play mind tricks on your user to steer them towards desirable behaviors.
If you can take something out without too many negative consequences, do it, because that will give you room later to make a completely different flow easier. Or leave that room empty to make your interface more simple and beautiful.
As for "scientific theories of design": Most of these things deal with visual weight and local presentation and have been common sense to artists for millenia, and are about as useful as A/B testing. That is, they are very useful towards the end to make sure that all of the separate elements in your design have reached local maxima. Maybe it's a bit like photography- if the photographer can give you scientific reasons for his white balance and exposure etc., good for him, but they say nothing about the subject matter or even the composition of the shot.
The ultimate criteria is the response of the audience, and no amount of tests or theories will ever touch that.