Bline people also want to communicate with sighted people, so it doesn't matter if they won't see it. You can also imagine priting the drawings out in a physical format they can feel.
We saw the Dragon MouseGrid years after coming up with this idea in IC2D. It is of course different in that Dragon's requires a visual interface, but it is indeed a quite similar idea (and that is more evidence it is a good idea). We have no problem crediting their work in future papers.
Others have tried similar ideas with musical/audio feedback. It is harder than you think. The results are quite mixed (you try using only sound to find the exact point again out of several hundred unique points).
Yes, I'm not a great sketcher (with or w/o computer). I cannot draw anything like the two figures you posted. I stand by my original comment. Also, the car was by someone who was blind from birth and has no experience drawing or using the program (IC2D).
Gestural input using a mouse or pen is not very new and has been around since at least the mid-80s (e.g., see work by Bill Buxton's group at Toronto). It could be argued that Sutherland's SketchPad used gestures way back in the early 60s.
The key advantage of gestures is that the user can specify a command and an operand all in the same motion. It takes less cognition than making a selection and then going to a menu/palette to select a command.
They are quite useful for pen-based devices, which have made a comeback in the last few years, but can be hard to learn for beginners. One innovation is marking menus, which pop-up a pie menu if you move the mouse slowly, but if you move quickly the pie menu never pops up and the resulting mouse movement is simply a gesture (for the expert).
We have created several research systems at Berkeley that use gestures (e.g., DENIM, a sketch-based web design tool.) We have also created tools and toolkits for creating gesture-based applications. Quill is a tool for training a popular gesture-recognition algorithm (by Rubine) that you could then easily incorporate in your own Java applications (our release includes Rubine's algorithm).
SATIN is a toolkit for creating pen-based applications and includes hooks for the recognizers and pen-aware widgets like pie-menus.
All of our code is open source, so feel free to download any of these systems or the source, try them out, and tell us what you think.
Slashdot Mirror
Bline people also want to communicate with sighted people, so it doesn't matter if they won't see it. You can also imagine priting the drawings out in a physical format they can feel.
We saw the Dragon MouseGrid years after coming up with this idea in IC2D. It is of course different in that Dragon's requires a visual interface, but it is indeed a quite similar idea (and that is more evidence it is a good idea). We have no problem crediting their work in future papers.
Others have tried similar ideas with musical/audio feedback. It is harder than you think. The results are quite mixed (you try using only sound to find the exact point again out of several hundred unique points).
Yes, I'm not a great sketcher (with or w/o computer).
I cannot draw anything like the two figures you posted. I stand by my original comment. Also, the car was by someone who was blind from birth and has no experience drawing or using the program (IC2D).
The key advantage of gestures is that the user can specify a command and an operand all in the same motion. It takes less cognition than making a selection and then going to a menu/palette to select a command.
They are quite useful for pen-based devices, which have made a comeback in the last few years, but can be hard to learn for beginners. One innovation is marking menus, which pop-up a pie menu if you move the mouse slowly, but if you move quickly the pie menu never pops up and the resulting mouse movement is simply a gesture (for the expert).
We have created several research systems at Berkeley that use gestures (e.g., DENIM, a sketch-based web design tool.) We have also created tools and toolkits for creating gesture-based applications. Quill is a tool for training a popular gesture-recognition algorithm (by Rubine) that you could then easily incorporate in your own Java applications (our release includes Rubine's algorithm). SATIN is a toolkit for creating pen-based applications and includes hooks for the recognizers and pen-aware widgets like pie-menus.
All of our code is open source, so feel free to download any of these systems or the source, try them out, and tell us what you think.
James