Not for what it's meant for, but how they used it is.
One of the main things you have to remember is that when people try to do something quickly they have to apply more force to move their limbs, which requires that you apply more force to stop your limbs as well. And getting high peak forces in place with just the right timing to stop perfectly on a specific point in space rapidly becomes more difficult the faster you're moving.
You can even see it if you look at people learning to draw a gun quickly to fire it at a target: they'll overshoot their target slightly and then settle on target. It's one of the reasons for the old saying 'Fast is fine, but accuracy is everything.'
I've digressed slightly, but the point is that it's the same for everything, including user interfaces: to do anything quickly people need either a ridiculously high degree of repetition, large hit boxes, or some sort of immediate feedback that they use to trigger a response that does something. (Like, you can type without looking at the keyboard fairly easily, but you can't type without looking at the keyboard on a tablet as easily - at least not without really good correction software - because it doesn't give you the same sort of feedback.)
Related, this is why the better shooter players often turn their mouse sensitivity way down. It's also why, when you're designing a user interface for a program you want the most commonly used options against the edge of the screen - since when someone flicks their mouse there they hit an effectively infinitely deep hit box. (Unless they're on a multi-monitor setup =p )
Stylus interfaces have pretty much always been designed around using the stylus to hit small targets - that's one of the reasons why they used it rather than your finger, so that they didn't have to re-write large portions of their interfaces. And when you do that you pretty much kill usability because there's little real tactile feedback to allow you to fine-tune what you're doing so you have to be looking at the device and taking your care over where you're placing the stylus. It takes forever.
To an extent, people did try to get around that sort of issue with handwriting recognition for text input. However, there's also an element of resistance that you get with pens and an even greater extent with pencils that you don't get with a stylus, which allows you to be more careless in how you apply force. Writing on tablets is messy and large, (and the translation of the writing into text is inexplicably bad.)
It's also worth mentioning that you can hold your arm much more loosely when you're not holding a stylus. As any half-decent boxer will tell you, tension makes you slow. So, again, that's an advantage for the no-stylus camp since you don't have to hold you arm in a particular way to keep the pen in the right position.
So, yeah. They're not bad, for what they're meant to do - mainly sketching. But to bypass not having an interface designed for touch, and to bypass having awful implementations of touch screens, they're the kiss of death. You can sketch with them, you can take (usually messy) freehand notes. But if you're going to be operating an interface entirely off them, then that interface is going to have to be large - and if it's going to be large fingers are better because you get better (though still not great) tactile feedback and you have that absence of tension.
And all this is when the stylus is good, mind, and doesn't require you to click on the icon for a bit to get it to register.
The type of stylus I was talking about involves an active digitizer and pressure sensitivity. You can draw and write pretty much naturally with them. Your handwriting with a stylus designed for a digitizer should be just as neat and fluid as your normal handwriting. The pictures you draw can be as rough or as finished as you want them to be. You are only limited by your own skill.
I agree that they are terrible for navigating an operating system.
One of the main things you have to remember is that when people try to do something quickly they have to apply more force to move their limbs, which requires that you apply more force to stop your limbs as well. And getting high peak forces in place with just the right timing to stop perfectly on a specific point in space rapidly becomes more difficult the faster you're moving.
You can even see it if you look at people learning to draw a gun quickly to fire it at a target: they'll overshoot their target slightly and then settle on target. It's one of the reasons for the old saying 'Fast is fine, but accuracy is everything.'
I've digressed slightly, but the point is that it's the same for everything, including user interfaces: to do anything quickly people need either a ridiculously high degree of repetition, large hit boxes, or some sort of immediate feedback that they use to trigger a response that does something. (Like, you can type without looking at the keyboard fairly easily, but you can't type without looking at the keyboard on a tablet as easily - at least not without really good correction software - because it doesn't give you the same sort of feedback.)
Related, this is why the better shooter players often turn their mouse sensitivity way down. It's also why, when you're designing a user interface for a program you want the most commonly used options against the edge of the screen - since when someone flicks their mouse there they hit an effectively infinitely deep hit box. (Unless they're on a multi-monitor setup =p )
Stylus interfaces have pretty much always been designed around using the stylus to hit small targets - that's one of the reasons why they used it rather than your finger, so that they didn't have to re-write large portions of their interfaces. And when you do that you pretty much kill usability because there's little real tactile feedback to allow you to fine-tune what you're doing so you have to be looking at the device and taking your care over where you're placing the stylus. It takes forever.
To an extent, people did try to get around that sort of issue with handwriting recognition for text input. However, there's also an element of resistance that you get with pens and an even greater extent with pencils that you don't get with a stylus, which allows you to be more careless in how you apply force. Writing on tablets is messy and large, (and the translation of the writing into text is inexplicably bad.)
It's also worth mentioning that you can hold your arm much more loosely when you're not holding a stylus. As any half-decent boxer will tell you, tension makes you slow. So, again, that's an advantage for the no-stylus camp since you don't have to hold you arm in a particular way to keep the pen in the right position.
So, yeah. They're not bad, for what they're meant to do - mainly sketching. But to bypass not having an interface designed for touch, and to bypass having awful implementations of touch screens, they're the kiss of death. You can sketch with them, you can take (usually messy) freehand notes. But if you're going to be operating an interface entirely off them, then that interface is going to have to be large - and if it's going to be large fingers are better because you get better (though still not great) tactile feedback and you have that absence of tension.
And all this is when the stylus is good, mind, and doesn't require you to click on the icon for a bit to get it to register.