Pens v. Mice
Last updated at 4:08 pm UTC on 14 January 2006
Some random thoughts on differences between Pens and Mice
I was thinking the other day about pen computers, and the difference between pens and mice. I have owned both an Apple Newton and a Palm Pilot, and have played with a WinCE handheld, and it occurred to me that there are differences between pens and mice, and their respective environments.
Two big differences struck me immediately: 1) A pen acts like a mouse with only one button: ie a Mac Mouse. 2) When a pen isn't touching the screen, ie when the mouse button isn't pressed, a system has no idea where the pointer is.
In the original Mac, the mouse was a good metaphor for a pen. There was only one button, and when the button wasn't pressed, the mouse did nothing except move the cursor around. Windows copied this interface originally, and for a long time (though most WinMice have two buttons) only used one mouse button. Recently, Windows mice appear to have been cross-bred with game controllers, and are growing all sorts or buttons, toggles and little scroll wheels. I recently saw a new mouse with 4 buttons and a wheel to scroll the screen. From the start, Smalltalk used three buttons on the mouse (does anyone know why? I think the reason is purely historical, the first smalltalk computers had three buttons, see: http://www.mucow.com/squeak-qref.html). A small number of buttons, where each button does a specific task seems to work well, but how do you translate this into a pen environment? A hack involving keystrokes, etc., seems arbitrary and messy.
When I take my pen off the screen, the system cannot track where it is. Again, in the original Mac the mouse acted as a metaphor for a pen: when the button was up, all that was left was a cursor to mark where the mouse was. The cursor did nothing. Nowdays rollovers, where a button self-highlights when a cursor moves over it, are becoming a staple of web designers and computer programmers. I like rollovers, they look cool and seem intuitive. But they don't work in a simple pen environment. Notably, Smalltalk seems to have had rollovers from the start: scroll bars that jumped out, etc., suggesting that when Smalltalk was designed, the mouse was seen as a unique device and not as a pen metaphor. How should rollovers be translated into PenSqueak?
One last thing. Showing system status by changing a cursor to an hourglass or wristwatch works poorly for pens, because when you take a pen off a screen, the cursor disapears. I'm not as sorry about this as I am about rollovers, because using the cursor to indicate system status is growing confusing in the age of multitasking. My browser, when loading a page, has a different cursor for each part of the screen: the page itself is an hourglass, the scrollbar is a pointer, the buttons are a hand etc. There must be better ways.
Well, here end my thoughts for now. Please add to this: it is a Wiki after all :-)
Bruce Tognazzini has said on his website:
"The Apple menu bar is a lot faster than menu bars in windows. Why? Because, since the menu bar lies on a screen edge, it has an infinite height. As a result, Mac users can just throw their mice toward the top of the screen with the assurance that it will never penetrate and disappear."
This advantage seems to me to go away with the advent of pen-based interfaces. The advantage is based on Fitts's law, which states that target acquisition time is in direct proportion to the distance and size of the target. With a mouse, menu bars pinned to the edge of the screen have "infinite size" because the focus is on the mouse cursor. With a pen, the focus is the eye, and every point on the screen becomes effectively equidistant.
This brings up another point: Fitts's law is based on the perceived size of a target. If a target is blinking or changing, wouldn't that increase the perceptual "size" of the target?
Allan, I think you misunderstood Tog's observation about the menu bar. The infinite size of the Mac menu bar has to do with the manipulation of the pointer, not with the menu bar itself. The menu titles are infinite in the sense that each one extends arbitrarily far off the top of the screen – once you've made it above the bottom of the menu bar, you can't push the mouse cursor too high on the Mac. For pens, this doesn't work because the pen can't be constrained by the software – it's easy to move the pen off the top of the screen, onto the screen's faceplate.
Anyhow, as for pens lacking buttons, I have two comments. First off, there are more sophisticated pens than what you see with your typical palmtop. The Wacom tablets, as I recall, can actually sense the pen before it made contact with the tablet, allowing you to physically hover over the tablet to see where you are on the screen. It is also a pressure-sensitive pen. Paint programs can use this determine the width and/or darkness of the brush stroke.
Second off, consider writing devices in the real world. There are the four-in-one pens (blue, black, red, and green) that let you switch colors by clicking the correct side of the pen. This is a pain in real life, and would probably be worse on a handheld, so I don't really recommend this. My second observation is the pencil eraser – like pencils, a computer pen could be double-ended. This would be a nuisance, too, but it would be a lot harder to get confused than a button on the pen would be (unless the button had a colored light – oy!).
Bill, I believe you've misread. If you'll note, the infinite height is a direct quote from Mr Tognazzini. As a Mac user since the spring of 1984, I am intimately familiar with its interface's advantages and shortcomings, and exactly how the mouse and menu work. And you repeat what I say earlier: the advantage goes away with pen-based interfaces.
As for your comment to Rusell, I have also worked with Wacom tablets, I find I don't mind buttons and switches on tablet styli so much, even find them useful. Though if you lose the pen on a Pilot, it's reassuring to know that you can use your fingernail without needing a cybernetic implant.
I'm neither familiar with the Mac nor with pens as input devices. While I find it very convient to have several buttons on a mouse (say one for selecting, pasting and context menues, repsectively) I can hardly imagine a pen with several buttons not to be akward. To address the question of an appropriate metaphor (which seemed to be one of the original purposes of this page), let me sum up what you said so far.
Touching the screen with a pen immediately translates into a left click. This means that drawing a line corresponds to clicking, keeping the button pressed and dragging the mouse. There is no need for a mouse pointer since it was invented to mimic the tip of a pen.
Now the problem is to simulate several buttons. (Did I understand everything correctly so far? Sorry for my English...)
What about having several virtual ink glasses? I envision dipping my pen into the context ink in order to open a context menu with the next touch of the screen. (Like changing the color of the next word by using red ink.) I'm not sure whether this does not dramatically increase the distance required to trigger a certain action but it somehow feels more "natural" to me.
Johannes, the Wacom Graphite tablet stylus has a long button along its side that can pivot, and an "eraser." Pivoting pointward activates one (assignable) function, such as right click, while pivoting eraserward activates another assignable function. Using the eraser usually erases, though that too is assignable. Not too difficult to handle once you get the hang of it, and using a muscle-memory mode (like a gas pedal) is supposedly a better user interface paradigm than the "real mode" that a virtual inkpot would be (Raskin?).