Just read (via DF) this 1989 article by Apple human interface guru Bruce "Tog" Tognazzini. In a nut:

We’ve done a cool $50 million of R & D on the Apple Human Interface. We discovered, among other things, two pertinent facts:

* Test subjects consistently report that keyboarding is faster than mousing.
* The stopwatch consistently proves mousing is faster than keyboarding.
This had a big impact on me. I've been a keyboard-shortcuts guy ever since my first job, where my boss would stand over my shoulder and correct me when she saw me reach for the mouse. Now the first thing I do in a new app is train myself to use the key commands, and I've created custom shortcuts in all the apps I use frequently (e.g. in Word, Command-Option-W for Word Count), and I use Quicksilver to launch apps, open files, search Google, send email, get lunch, basically everything. All of this keyboarding makes me feel very efficient. And now here's Tog himself bringing my world crashing down around me.

But when you think about it, it can't be as simple as Tog suggests. The blanket statement, "Mousing is faster than keyboarding" is, presumably, true in certain circumstances. But it can't be true always and everywhere.

I spend a lot of time writing in Word. (I know, I know, but I'm used to it.) I try to write 2,000 words a day, and although I don't always manage it I usually get close enough. Based on a random sample of my prose, that's about 10,865 characters. I enter almost every one of these 10,865 characters into a Word document using the keyboard. According to Tog I should be able to save time by finding the character in the Symbol dialogue box and clicking on it.

Maybe that's a facile example. Tog might say, "Of course, I didn't mean typing words. That's what a keyboard is for. I meant performing other actions."

So here's an example that's more on point: saving. While I'm writing my 2,000 words, I am a saving freak. I save my document reflexively. Whenever I'm not typing, I'm saving. I'm sure I take saving to a useless and neurotic extreme, but it's a harmless neurosis -- the computer can handle all that saving, and it removes a source of worry, and I never have those I just lost two hours' work things that happen to other people.

I do all this saving using the venerable Command-S. I did it just now, after typing that last sentence, autonomically: hands in the resting position, left thumb about an inch to the left (I'm left-handed), left ring finger down. Boom, saved. Not once do I think about the Command key or the S key, just as I don't think about the Shift key or the T key when I begin to type Tog.

I could, instead, use the mouse to go to the File menu's Save command, or to the Save button in the toolbar. I find it hard to believe that would be quicker, but perhaps I'm falling victim to Tog's first point and failing to accurately register the time it takes to hit Command-S. So let's abstractify a little. I can't say for sure how fast I am at hitting Command-S, but I'm definitely faster than I was when I started using a computer. I'm faster than the average computer user, just because I do it so often. Either the speed of mouse-saving is like the speed of light, and there's no way you can ever catch up with it, or at some point I'm going to be faster with the keyboard than with the mouse.

Abstractify one layer further: If a keyboard shortcut is used frequently enough, and the buttons used are convenient and memorable enough, and the mouse alternative is sufficiently complex (identify the Save button from all the buttons on the toolbar, find the cursor, land the cursor on the Save button, click, return hands to the keyboard), then the keyboard shortcut is quicker and less distracting. If I only saved once a day, and the shortcut was Control-Option-Y instead of Command-S, and Microsoft had made the Save toolbar button twice as big, and there were no other buttons next to it on the toolbar, then using the mouse would be quicker.

And what about more data-dense applications? When I'm editing audio in Pro Tools and I need to move my cursor to a particular spot, there are 44,100 possible cursor locations per second of audio and maybe five minutes of audio represented on my screen. I can try to find that spot with the mouse, using repeated clicks of the zoom button, recentering, squinting at the waveforms, then unzooming back to the original view. Or I can hit the Tab key and, using Pro Tools's Tab to Transient feature, allow the software to find the exact spot I need. Is the mouse quicker then?

Tog wrote his piece in 1989, the year the first version of Pro Tools (then known as Sound Tools, which is a better name) was released. He can't be blamed for not knowing about high-resolution audio or video editing. Still, one wonders about the $50 million worth of testing he did. Did he test on anyone who'd spent ten years hitting Command-S as often as I do?

In fact, the answer to the question Which is faster, keyboard or mouse? is not Tog's one-size-fits-all answer (the mouse, and testing proves it!), nor the answer of my old boss (the keyboard, and get your hands off that mouse!). It's For what user, attempting to accomplish what task, under what circumstances?

Update: This 2005 paper (PDF) from the International Journal of Human-Computer Interaction comes down squarely on the side of keyboard shortcuts.