Adaptive Interfaces: an introduction
Interface design is complex. Designing a user interface which can be used efficiently by a large group of users, preferably without any form of training is even harder. Things get even tougher when these groups of people have different tasks to accomplish. Are we splitting up the screen across different views or pages? Or do we overwhelm the users with a gazzilion buttons?
In a perfect world, interfaces adjust for a given task. Touch screens are a perfect example. Instead of a keypad and navigation buttons, the buttons are laid out differently for each task, while providing a nice piece of screen real estate. The mapping between an element and action suddenly become way more clear.
When looking at devices with limited screen size, we often see interfaces which are highly focussed and have a low level of complexity. You can only perform a few tasks, which makes completing a task way easier. This is often why mobile applications are often percieved as easy to use. The downside is that often you need to drill down several layers of menus to find a certain functionality.
So far we've established two things:
Limited interfaces are easy to use
In order to have limited interfaces, you need to have limited functionality or use a lot of layering to deconstruct tasks into simpeler tasks.
Imagine a world which doesn't require this kind of surgically deconstructing. Welcome to adaptive interfaces.
Tradionally, interfaces have been static. They do adjust when the user flips a certain switch, but every time you enter a certain screen, you start from the beginning. Adaptive interfaces are dynamic. The interface changes over time, powered by advanced machine learning algorithms.
If you have a hard time grasping this principle, think of the concept as an advanced version of the iGoogle homepage. The pages are build up of several widgets, which can be added and moved around as the user desires. Ideally, adding, removing and reorderig of the widgets is done entirely automatically.
Adaptive interfaces aren't widely used, yet. They are tough to implement and hasn't seen as much research as the static interface. I expect these this concept to take off within a few years; the technology is ready, the developer are ready and the users are ready. I can't wait to see these UIs pop up in the wild.