Nathan – Peter - Kaitlyn
Initial Design Sketches
Started with Crossing Cross over, cross back, etc Gestures Acceleration/Velocity based Select
Wright Cursor
Pure crossing seems to be: Error-prone Unusable in current interfaces Unsafe
So we went to : EDGES!
Edges: Very accurate Still Crossing, just like Mac Menubar Pretty Safe Ah Ha! – Make everything as easy to click as the buttons on a screen edge!
Flipbooks: Hash out our Ideas We also tried a binary search, quadrant area cursor. While it has potential, we think edges are too powerful to pass up.
An area cursor follows the mouse, separating the selection area into 4 quadrants
Activation causes the cursor to move independently of the area cursor.
6 4 Moving to the edge of a quadrant modifies the selection to contain only the items from that quadrant. The edge constrains the mouse and keeps it in the area cursor
The items from the selected quadrant are rearranged into quadrants. Now, going into a quadrant selects the remaining items. If more than one item is in the quadrant, the process is repeated
2 6
2 Selecting a quadrant with only one remaining target selects the target.
Target Activate tunnels Keystroke Click
Target What if the mouse is over a pliant region? … that ISN’T your target?
Target What if the mouse is over a pliant region? … that IS your target?
Target What about right-clicking? Open Copy Properties
Edge.Point
Interaction Time!
Pros: Plausible! Does not hamper user who do not need or want pointing assistance. Works with current desktop interfaces. Effective widths of targets = ∞ Steering tasks are automated by tunnels. Resize area cursor via mouse-wheel.
Cons: Must deploy area cursor to gain assistive benefit. Mouse buttons are used. Algorithms could be improved to better map tunnels to the cursor.
User Tests
Without explanation or practice, the technique was very hard to use for our subjects. They ended up just pointing and clicking the icons. Normal interaction still works perfectly, a feature of our system.