I was recently shown this collaborative editor called Gobby (there are others) and was reminded of an idea I’ve been toying with for a long time.

A lot of work these days goes into novel human-computer interfaces (think coffee table displays, networked whiteboards, etc.) and gestures (think various touch responses like multi-touch zooming), but in my opinion (and metaphor), these work are like lightly scratching the skin when there is a deep deep itch.

The deep itch is a totally multi-input environment. I don’t mean multiple non-interacting or minimally interacting users on the same machine. I don’t even mean multiple users who interact only when it is obvious to do so, like during conferencing, or when the game players are in one area of the map. These are all fundamentally still single input environments, because each user owns a static token, be it the single identity in a conference, the single drawing pen on the virtual whiteboard, or the same virtual character in a game!

Why not be bolder? Suppose two people are working on a desktop, whether locally or via a network connection. Why not have two mouse cursors and two possible focuses? It works like this: any place that cursor X clicks receives focus for all of cursor X’s associated inputs (perhaps keyboard X). Each user gets a set of inputs, then they can be working on separate windows, or even separate parts of the same window. They can both see what is going on with the both of them so they can really collaborate. Change two users to some other number and the same applies. None of this VNC-like fighting cursor crap. Let each user have multiple sets of inputs and still the same applies. See, now you can really multitask with two hands and feet if you want rather than glorified time sharing — lots more possibilities here than boring gestures! It’s how the real world works anyway. Besides, if processes can share a processor, users can certainly share a virtual interface and all of its objects and resources.

It seems like all — and I do mean all — interfaces existing and under development suffer from some kind of designer brain freeze where each user is assumed tied to a single and static virtual token in the interface, simply due to the coincidence that the user is tied to a single and static physical input set. That’s stupid. If you could have 10 hands rather than 2, wouldn’t you rather have 10 hands? Well, in the virtual interface, you can and should. And if your 10 hands can all work, and your friend’s 10 hands can work, you should be able to work together so you have 20 hands in total. Swap hands, too.

This is so obvious, I don’t even understand why it hasn’t been done 10 or 15 years ago…