Touch versus pointerApple versus Microsoft (Windows 8):
- Apple. Right: “People don’t touch their desktop screens”. Wrong: You must touch your screen when using your iPad with a Bluetooth keyboard.
- Microsoft. Right: Support both touch and pointer modes. Wrong: “People will want to use touch everywhere”.
Phone versus tablet versus desktopApple versus Microsoft:
- Apple: Right: Tablet user interfaces must be simpler than desktop user interfaces. Wrong: You need to completely reimplement a the user interface when going from phone to tablet to desktop.
- Microsoft: Right: Striving for reduced complexity makes sense on desktops (and not just on tablets). Wrong: People want the complete desktop experience on a tablet.
- Small (cell phones): minimal complexity, support the most obvious operations.
- Medium (tablets): moderate complexity, support 80% of the operations that users need.
- Large (desktops, notebooks): high complexity, support all conceivable operations.
How to implement cross-level user interfaces. That is a tricky one. Note that desktops can profit from a “simple mode” (tablet-like). Optionally, one switches to a an “advanced mode”. If we can figure out how to do that properly then my preferred solution would be to start with a simple mode and incrementally expose advanced features, on demand. Ideas for user interfaces that adapt their complexity to the current screen size are already out there:
- Enyo allows webOS apps to scale their UIs.
- Fragments are a mechanism in Android for splitting up a user interface so that the single fragments fit on cell phone screens, while multiple fragments can be shown on tablets.
OutlookA few more thoughts:
- This post ignored “couch user interfaces”  where your screen is much farther away and you often have a remote control.
- Two means of input were also completely ignored: gestures (Kinect) and voice (Siri). These have many applications, especially for couch user interfaces.