The original note, as written at the time
How about this for an interesting new peripheral input system? Here you have a mouse. It’s a wired mouse, but along the cable, which is quite long, are sections you can describe as nodes. Each node has a button. The first button is for microphone input. The second button is for spacebar input. The third is enter. The fourth is copy, and this could be a hybrid for copy and paste.
Ongoing development
A peripheral input idea: a wired mouse whose cable carries a series of nodes, each with a button.
The first is microphone input. The second is spacebar. The third is enter. The fourth is copy, and it could be a hybrid for copy and paste.
What makes this good is that it puts the most-used keys of a physical workflow onto the hand that is already resting somewhere. Everything you do while writing happens at the keyboard, and the mouse hand is idle during it — or, worse, it has left the keyboard entirely to reach for something.
The key that earns its place is the microphone. Not as a novelty: push-to-talk on the mouse, held down with the thumb while the other hand stays on the keys, is the correct interface for dictation in a real workflow. Every current implementation asks you to press a key on the keyboard, which means stopping the thing you were doing.
Spacebar and enter on the cable are more questionable, because they are already under the other hand. The honest design set would test which four functions a person actually reaches for while drafting, and expect the answer to differ by profession — a designer wants undo and zoom, a writer wants copy and the microphone, a coder wants bracket and semicolon.
The failure mode is a cable covered in buttons that you press by accident. So the nodes need to be placed where the hand naturally pauses, above the desk, not along the whole length.
Nothing here is a new technology. It is a rearrangement of what exists, which is why it is a good idea: it costs almost nothing and it can be tested with tape.