So having created this mechanical design that could function, there was still another component of the user interface that I worked on and David Kelley worked on: which was not only what it should look like, but what it should feel like. What would it feel like in your hand, and how would you interact with it? Apple worked a lot on the one, two, or three-button mouse; Xerox had a three-button mouse.
What we immediately noticed-- not knowing what a mouse was supposed to do-- was that it was hard remember which finger was supposed to push which button to do what function. So the Apple mouse went from having three, to two, to one buttons-- which is another significant contribution that Apple made to computing, the one-button mouse.
A sidebar for a second. It's also interesting to note that in 1980, at Hovey-Kelley we didn't have access to a computer to plug a mouse into. So we were a little puzzled as to what this thing really was going to do. People described it: they said, "Oh, it's going to be the primary interface for the computer," which we thought was laughable-- you could balance a checkbook without a mouse, and you could write BASIC programs without a mouse-- that was about all people were buying Apple II's for. That and playing games, and there were game paddles and joysticks, or keyboard input, and nobody thought of what a mouse would be used for in a game. So we had to build an interface card for a mouse for an Apple II, because none existed, and then we had to write software to keep track of the position.
As it turned out, the computer spent most of its time just keeping track of the mouse. Our great accomplishment was to move a dot around on a screen. We wanted to make sure that if you drew circles with the mouse, it wouldn't wander off the screen-- there'd have to be an absolute positioning so you didn't have to constantly pick up the mouse. This is something that Doug Engelbart's mouse had a problem with: it would drift horribly. He pointed out in his demonstration that you could simply pick up the mouse and move it over to re-zero it. We felt that was unacceptable, that it had to be accurate enough that you could move it around for a long, long time without repositioning it.
So it was laughable that we were designing the mouse of today by moving a dot around on a screen. There were no pull-down menus, there were no cursors, there was no highlighting. We only saw a demonstration of a Xerox PARC mouse plugged into a Lisa breadboard computer, attached to a couple pieces of plywood. They had a couple pull-down menus, and they could change fonts, but it wasn't something you could really exercise. That came after the mouse was designed, interestingly enough.
(Prototype mice shapes)
So one of the big debates about the early mouse was, Was this something delicate that you would hold with the tips of your fingers, or was this something that you would grab, like the stick shift of a car or a sanding block?
We made probably a hundred different shaped blocks; David Kelley probably still has a box of them in his office. I think he cut the ball off the stick shift of his BMW, which was a dimpled ball, and turned it into a mouse. He made a sanding block into a mouse. We had handles off of bicycles, we had all kinds of odd-shaped things, things that were molded to your hand, things that were perfectly rectangular, to see what the right design was.
The original mouse was designed for the Lisa-- the Lisa came first. The end result was something that we felt you didn't want to grab; you wanted to hold it somewhat delicately, and have your index finger free to click the button. So we ended up having a gentle curved shape, the size of a pack of cigarettes, roughly, and followed the industrial design of the Lisa keyboard-- which was quite a large, substantial keyboard compared to keyboards today. And of course the mouse has evolved over time to the current rounded mouse shapes. And for a time there were mice that were more like sanding blocks; even Logitech built mice that were very very large. But I think in general that, because you want to move your hand to the mouse and back in a gentle motion, not an aggressive, grabbing stick-shift motion, that the smaller designs have become more prevalent.
As for the one-, two-, or three-button design, we just built all kinds of different prototypes of buttons, and one button seemed easier to use. It might have gone differently if there was software to use that would have shown why you needed a second button. At Apple, the evidence that the mouse was designed first, rather than the user interface, is that only later on did the idea of option-clicking, shift-clicking, command-clicking come about. If we had known that that was required, it might have forced us to make a two-button mouse. I'm happy that that didn't happen, because I think more young children, and more people on planet Earth, have picked up computing because knowing that there's only one button to click. It just makes the learning process that much simpler.
Another characteristic which I think is prevalent among mice today is, despite the fact that we'd made a mouse that was compliant with dirt, we were concerned that dirt could still get inside. So one of the first things we did was build a little wiper ring that would wipe the dirt off before it could get inside; but that tended cause too much resistance and was quite noisy. And we ended up coming up with a nifty ring interlock. The user could take it off without any tools, pop the ball out, and expose enough of the parts so they could be cleaned. Any mouse today has essentially the same design. That was a James Yurchenco design: to twist it off, pull the ball out, clean it, and put the ball back in.