Slashdot Mirror
Could a Pen Replace the Keyboard?
theluckman writes "Reuters has this story on how new devices like "digital pens" could possibly replace keyboards as primary data entry devices. Maybe so, but I would need my pen to make cool clicking sounds."
← Back to Stories (view on slashdot.org)
Geez. I wonder if I spend too much time on my handwriting. I clocked it with a stopwatch once on "the quick red fox..."
and I could barely hit 20wpm cursive, a little slower printed.
I can type at 90-95.
Incidentally, I use a trackball that I hacked to have an external box with the mouse buttons so I can operate the trackball with my right foot and the buttons with my left (due to RSI). So I never take my hands off the keyboard. I can't see myself going back to having to use a hand mouse, pen, or whatever as my primary pointing device.
Robotiq.com is heavily tested on animals
If the stylus had ever been faster and more efficient than those "lowly keyboards not so different from the ones that powered the Smith Coronas and Ollivettis of yesterday," then nobody would have bothered with those Smith Coronas and Ollivettis in the first place.
I'm not sure why the article starts by making fun of the venerable keyboard, since it serves such a different purpose.
Now, if you told me that this laser pen might replace the mouse (which, in fact, the rest of the article seems to do), that would be a different story. It seems to me that a pen could do everything that a mouse can, and, in most respects, do it better.
Accountability on the heads of the powerful.
Power in the hands of the accountable.
This is ridiculous, and the article on Reuters reads like a press release not an objective article. About what you expect because it looks like someone's just paid to have it stuck on the wire, but hey, worth remembering that.
:-)
With a pen I'm using three fingers to perform data input, with a keyboard I'm using ten. Far more efficient resource usage, and any character can be made with a simple twitch of no more than two fingers, while I line up the rest of my hand for the next characters.
I can't write anywhere near 100+WPM with any legibility, but I can type at that speed with pretty good accuracy (and I'm not exactly unusual...).
Think back to exams, and 5-10 pages of handwritten text in 2-3 hours. Major cramp problems, which I simply don't get producing way more input than that with a keyboard.
Replace joysticks? Come on guys, I've used a pen on a touchscreen as a joystick replacement before, it's woeful. Replacing eyeball tracking cameras as a data input system? Well, if anyone can come up with an example of someone who's physically capable of gripping a pen but who makes any quantity of input by this method, I'm amazed. Put simply, that claim is extraordinary enough that I demand a reference.
PDAs and phones? Well, most PDAs have touchscreens already so don't need anything this complex unless people want to input text to them by drawing on another surface, which seems to miss the point of a portable device. Phones? Cheap, commodity things with little data input that have to be rugged and survive teenagers? The pen makes them expensive and is going to get lost _really_ quickly. And who needs it, exactly? I mean, with decent predictive text we can already write at a pretty good speed for the length of input.
A pen is nice for drawing, some people like them for GUI use. Personally I like a touchpad which I can use without significantly moving my hands from the keyboard but hey, everyone's different
Someone has had a bright idea and has oversold a story to Reuters, who've published it straight. No problem with that, they're a wire service not a newspaper, but this isn't a credible story. These people aren't going to take over the world and their claims are rubbish.
Greg
(Inside a nuclear plant)
Aaaarrrggh! Run! The canary has mutated!
Keyboard are better, given. BUT, for mobility, they just suck. It would be much easier to push a button on your pen and scribble a text message on the back of your cell phone than txt w/ lot sht-hnd. This can also be practical for automagically transforming notes to your computer. Depending on how much ram is in there you could possibly scribble someone's phone number on your hand without actually using ink. At a meeting you wouldn't have to carry the current bulky electronic clipboards, just your pen and maybe an extra memory stick.
Keyboards will always rock the desktop input world...until we get neural implants. (:
--Roy
The average speed of talking is 125wpm, while the average speed of writing is 55wpm.
The thing that works very well for me, I can type around 140wpm (All Hail Mavis Beacon) so talking has become a mostly inefficient method of communication. However, I've found most of the time while speaking I say something stupid I wish I wouldn't have said -- while typing something out I can at least backspace it out. Amen for the backspace.
Dacels Jewelers can't be trusted.
Well, what if you could program the syntax into the pen? You could write code on smart paper on the train or anywhere, and have the pen bebug as you write? That would be handy.
keyboards are great...but only for english and other languages that use the same character set. But when you have to deal with other languages like sanscrit or chinese, most people probably write them faster than they could type.
most character/pen recognition systems kinda suck. But i've heard, and my friend is trying, that a program called atok works pretty well for pen recognition in japanese for the palm pilot.
I think that's the more interesting question. Keyboards can handle 100 wpm, and there's no filter between the process of writing a character and the computer interpreting the character. Pen and voice would both need this.
But even assume the computers are so fast that there's no slowdown with that interpretation. Will it ever be easy to manipulate a UI with your voice? Possibly for some things, but about more content-focus software like a word processor? How about if you're writing a manual about how to manipulate a UI? Could you imagine the amount of escape characters at work in your dialog at that point?
And even if that first draft was easy enough to do, how about all subsequent drafts? "Computer, go to line 135 and replace the second occurence of 'there' with 'their', and that's 't - h - e - I - r'." Sounds a little clunky to me...
But I suppose the folks at Dragon et al. have already run into these issues and found solutions for them...
--------
Bleah! Heh heh heh... BLEAH BLEAH!!! Ha ha ha ha...
One of the reasons why the Apple Newton PDA failed so miserably was its promise of usable handwriting recognition. Unfortunately, that promise turned out to be more a case of wishful thinking.
Heh. Good thing you're not modded up, or you'd be perpetuating this myth- one that is especially blindly spouted here on slashdot.
The first few models of the Newton's HWR sucked. Pretty bad. After a year and a half, Newton OS 2.0 came out, with new HWR recognizers, and it got it right. Far faster input than Graffiti or other character and stroke based methods.
Fortunately, real HWR didn't die with the Newton. ParaGraph's CalliGrapher exists for WinCE, providing a more efficient, real HWR based, means of inputting on a PDA. There is also a version for the Windoze on the desktop called PenOffice. Unfortunately there is no such thing as real HWR for the Mac or Linux platforms though.
Having used both a Newton 2100u and an iPAQ with CalliGrapher, both a Newton and Palm device with Graffiti (originated on the Newton), Jot, the built-in character recognizer in PocketPC, as well as various programmable character recognition means, I've quite a bit of experience with HWR in the real world.
It appears that you don't have experience with much in the way of HWR, except perhaps on a Palm. That's fine, but it isn't very scientific to pull stuff out of your rear without any
experience to back it.
After 3 months of using my Newton and iPAQ (w/ CalliGrapher), I found I can get between 40-60 WPM. That was not counting any words fewer than three characters, so that number may be higher, but I wasn't sure how to determine WPM for sure. That's including making corrections. Around 99% accuracy for words, 90% for punctuation.
I tend to get higher WPMs on the Newton, mostly because the larger screen accomdates more words at a time, and that the recognition is rolling, rather than happening at once when I lift the pen. That is, if I write "hello my name is armondo," it will have recognized as text "hello my name" by the time I am writing the word "armondo."
Individual handwriting recognition technology for the masses may still be a pipe dream.
Try a real HWR system for a while, meaning a month or two. The same amount of time is required to get used to Graffiti, so I think that's fair. During that time, correct it. The real HWR schemes of which I know train a neural net against your corrections, and learn your HWR style over time.
Working toward a usable PDA environment in the spirit of Newton OS: Dynapad
That's where you are wrong. OCR might be more difficult, but this is not OCR. That's even a bit of what allows grafitti to work. The whole point is that it's recognition of the drawing of a character.
In those 'harder' languages, the people are very touchy when it comes to writing the language. Each of those complex characters has an exact number of strokes, with the order and even direction exactingly specified. Given all that, recognition of Kanji characters turns out to be much easier than of English characters (just think of how many ways one can draw the lower-case letter 'a').
That's one of the reasons that PDA have been a huge success in Japan. The Sharp Zarus line has been huge over there, due much in part to their successful Kanji recognition.
One could almost argue that grafiti is a success exactly because it applied the order of Asian language writing onto English characters.
Kids coming out of schools will soon all have better typing skills than writing skills - there's no reason to make all new equipment use pens and slow down input to computers. Even a crappy typist that types 30wpm cant match that with a pen, not for extended periods of time. I can type for hours with no problem, but I remember writing exams being quite painful.
We need increased speed of input devices to computers, not pens.
Hell, we should be criticizing the keyboard for its short-sighted 'one key at a time' input and go to a chord system which some people have gotten up to 200wpm on on custom versions.
And there's no reason to have a pen when keyboards
can now be projected onto a surface according to a recent slashdot article....
Maybe its not that big an issue, but I know that myself and everyone else who has suffered through 2 hour long essay tests finds that their writing hand is in a great deal of pain. I don't have this problem with a decent keyboard.
:,;,.,,
Also there's another issue I hav't seen mentioned. Unless the pen functions as a mouse as well, you will either have to learn to use a mouse with your off hand or switch back and forth.
If it doubles as a mouse, would that mean you'd be tapping the pen against something non stop while playing quake? That would require a lot more muscle movement than a mouse finger click. It kind of reminds me of when everyone thought touch screens would be a great idea until they discovered Gorilla Arm.
How do you tell the difference between characters like this:
", ', |, l, 1, `, \
-, _,
(, [
You could probably get some of the above using context but that will only get you so far.