This looks way too hard to get used to initially. The best thing about Swype (which is truly remarkable) is that you already know how to get to all of the letters. You dont have to relearn anything. You're able to rapidly type things out of the box, with relatively low error rates for a mobile keypad, without even thinking about it.
Texting on a number pad looked too hard to get used to initially, but now they hold competitions for speed on them.
I don't see this overtaking the entire industry, but it will at least make us think and try new things to come up with something better than a basic on screen keyboard.
Texting on a number pad never looked hard to get used to, it just looked slow (and still is). They wouldn't hold speed competitions for it if it was fast.
T9 drastically improves typing speed on numpads. I could definitely type faster on my old Nokia 6310 than I can on my iPhone. I've only had my iPhone for a year but you think that'd be long enough to learn how to type on it.
One thing I've found helps a lot is to just allow yourself to finish a word with a mistake in it and hit space, assuming the auto-correct will get it right. It feels backwards at first, but I think "scrolling" or "backspacing" through text requires a lot more of your attention/a context switch than just correcting at the end of a word.
I guess that does seem a backwards to me, probably because I don't type all that much on my iPod touch. My inclination has been to wish there was an option for the autocorrection to be a little less auto, offering a suggestion but assuming I meant what I typed and requiring me to explicitly accept the correction. If I did more text entry on the device, though, I could see myself coming around on the matter.
When I try to type quickly not even the iPhone knows what the hell I meant to write. :)
And that doesn't help for words it doesn't know. Turning auto-correct off might actually be a good idea for me. I'll have to try it and see if I like it any better.
T9 was incredible; it's one of my favorite pieces of technology. That said, I had a Palm Treo for a long time and that is my all-time favorite typing experience on a mobile device. The slightly rounded keys allowed great accuracy and tactile feedback (both for the "click" and for sensing where the correct keys were without looking).
I have now had an iPhone since the first generation and I still suck at typing on one (on the 3GS now). Auto-correct seems to hurt me more than help even after earnest efforts to trust it and stop hitting backspace.
Swype is extremely resource-intensive-- too much for older phones (such as my MyTouch 3G). If 8pen is less of a hog, and it seems like it probably would be, since it doesn't have to guess at what you wanted to type, it will be a great addition to the alternative input landscape.
My eris works great with its default settings...until I run a music player while trying to type. It would be great if they could dynamically scale the accuracy or something to keep response time up.
Compare relatively low error rates to zero error rate.
And no matter how good you get, no one (very few people?) can use Swype accurately without looking.
And intuitively speaking, it seems to me that if they've truly got some science behind the placement of the letters, it could be very, very easy for motor memory to catch on to this.
I worry that this will be the biggest obstacle for 8pen to overcome. I'm planning on trying it out, for sure. I think the learning curve will be steep, but if you can delay gratification it'll be a pretty nifty tool.
I agree that a tutorial would be useful, and that's exactly the problem.
From trial and error, people build up gazillion mental models of how certain things work. When I show you a bottle of red wine, you most likely know by glancing at it, how the bottle can be opened. That is because you've built up a mental model of bottle mechanics as well as different type of bottle caps and can map them to the new problem space (new unseen bottle).
When inventing something new, especially something as basic and "known" as an input method, you can greatly ease adoption by relying on existing mental models of things people already know, then map your new interface to that.
Swype would probably work reasonably well without displaying the underlying keyboard, but I assume they do so to guide the mapping process.
Recently acquired BlindType does the same. I remember that they even state somewhere on their website, that the typing works without displaying a keyboard at all. Unfortunately the site is down after the google acquisition, but this youtube video illustrates the "typing without keyboard": http://www.youtube.com/watch?v=m6eKm1gUnTE (skip forward to 1:02)
As someone who hasn't used either method, I disagree that this 8pen system looks hard to use:
-stick your finger in the middle dot
-move it into the quadrant where you see the letter you want to type
-move your finger toward the letter you want to type and keep circling until the letter is bold.
- back to center
I hate stumbling over autocorrection and text prediction features (which I assume swipe relies on), so would definitely choose 8pen if I had to make a choice based on the demos I've just watched.
The reason Swype is so easy is that you don't really have to look to see where the letters are, it's QWERTY, you already know. You can type blind on it after a tiny amount of use once your brain has mapped out the keyboard size.
I imagine you'd be able to use 8pen without looking too, maybe even better after a long while, it would just take you a lot longer to get there.
What is frustrating about this is that I would gladly pay $10-$20 for Swype on my phone, and I have never bought another app. They must be making a lot from preloads, or the preload agreements stipulate that they cannot sell the app through other channels, because everyone I know with an Android phone would do the same. Seems like a lot of lost opportunity.
There are also a bunch of swype apk's floating around that you can just install. Check the xda forums for your phone model, there's bound to be people who found one that works for it
It would seem they designed their site to look like something Google would make. The colors used in the lines that separate the quadrants on the 8pen are also reminiscent of Google's logo. Really hoping for that acquisition?
I was thinking the exact thing. Even watching the video it seems like they really tried to copy Google's brand... not sure if I can trust anyone that blatantly copies someone's brand.
It doesn't really offer advantages on a computer that it does on a small device. Full size keyboards already allow no-look, potentially 100% accuracy typing, with multiple fingers at once. That spanks this system even worse than this system spanks mobile phone keypads.
Just bought it, it's VERY difficult to use, even harder than I expected.
There's no control over the insertion of spaces, it automatically inserts one after you release your thumb at the end of a word. So if you accidentally delete a space that you need, which is likely during the learning process, there is no way to get it back without also deleting the character before it and adding an auto space. Also there is no way of preventing this automatic space either, so web addresses and the like have to be entered as one very long continuous gesture, without any mistakes; if you make a mistake you must release your thumb to hit the backspace button, which you must now hit twice because an unwanted space was just added.
Finally, the learning process is heavily hindered by the fact that your thumb obscures at least one or two of the 'branches' during a gesture, which means when you have to visually consult the screen because you have forgotten the unusual spatial model (the direction matters too remember,) your only option is to pivot your thumb without it leaving the screen, and hold your phone at an odd angle so your can peer beneath your digit and decide where it should go next. Not very practical.
Give me a spacebar and the ability to disable autospacing and I'll try it again. Until then I'm sticking with swype, which allows me to type extremely fast.
You can avoid entering spaces by lifting your finger when it isn't touching the centre area. It disrupts the flow significantly when you do so but it does let you stop half way through a word without putting in an extra space.
After playing with it for a while I think that it's an interesting idea. I'll stick with it until I'm familiar enough to make a fair judgement but it's difficult to imagine 8pen being as fast as the standard keyboard - there's just too much movement involved for every character.
Just got it, too. Had a short conversation via IM with it, and I'm already 100% faster. I agree with you on the spacing issue - I wish that you had to tap without a gesture on the center point to insert a spacebar instead of automatically inserting one. That way I could consult between letters.
Eventually muscle memory will kick in.
I'm forcing myself to use it for a week, and will compare typing speeds then...
The 'don't end on the center to lift and then continue the word' behaviour is invaluable to me as I learn, and I have no idea why it's not mentioned in the tutorial.
Also, if you make a mistake, you can go back segments as well as forwards as long as you don't touch the center. Useful if you go to far or (my favourite trick at the moment) realise that you've started in the wrong direction...
This is good because it's spatial. The last really spatial typing method is the traditional number pad. And it's reaaaally good despite the necessity of having to press each button several times most of the time.
A spatial method can be applied without seeing the screen. The speed and ease of spatial methods comes from the deterministic and predictable behaviour: you learn physical finger patterns instead of reading the screen and adjusting yourself all the time. So, in other words, it's one-way input instead of two-way interactive typing.
If the screen surface had a tiny little dot embossed to the center, blind writing would be even easier. The 8pen looks a bit weird though, so people might have a hard time figuring it out first. If the quadrants were aligned to 0, 90, 180, and 270 degrees it might be easier to grasp. Also, I hope the center isn't mandatory: you could just start from quadrant n and move over to quadrant n+1, lift your finger, and start over.
Keyboards are nice because after a while the location of each key is encoded in your brain. Swype and its ilk (ShapeWriter, SlideIT) take advantage of that by piggybacking on your already learned spatial knowledge.
This, however, is gestural rather than spatial: each letter is encoded via an action, rather than a point in space. What this means is that there is the same cap on efficiency as with handwriting.
Perhaps the best marriage of what 8pen is trying to do is basically something like Swype, but on a keyboard layout that's optimized for the cramped space. The pre-encoded spatial information will be sacrificed for greater eventual efficiency. You would start out tapping, then graduate to the spatial gestures.
(Re: the center, yes it's mandatory. They use non-center gestures as macro storage)
For what I saw, the input screen always seems to be the same and in the same location, thus the input method is spatial. That is, you learn approxximately how far each segment is from your handgrip and you can then whirl your finger around the correct quadrants to produce letters, even without looking. Pretty much how you operate a keyboard once you correctly learn the locations of letters.
A gestural interface would allow you to do, well, gestures. A gesture is like a vector: it has no location. You might learn that 'e' comes from "starting from top, two segments to the right" but that gesture wouldn't have a meaning unless you start from the correct location in the screen space.
On the keyboard front, we'll have to see. Swype, Graffiti, and the old fashioned keyboard are more intuitive because they borrow something core about our usage of inputting stuff: how we hand write a character for Graffiti, and everything we know about a QWERTY keyboard layout for Swype. Swype is very easy to grasp. For some reason, I see a much more competitive 8pen as a gesture/menu command issuing system than a keyboard. Replace the characters with commands and you have a powerful widget, even for games possibly. It isn't a random set of umpteen gestures, there is some hierarchy and precision built into the thing.
I feel the keyboard domain doesn't make use of hierarchy properly though. Let's say the 8pen keyboard divides the keys into octants. Each octant does not have a collection of related things. The letters I D G Z are not related in any way, yet they are all arrived to by first moving up and rotating right.
To illustrate a better use of the hierarchy, I'll divide the thing into quadrants instead of octants. Consider a widget to edit font. The top quadrant could be font family, the left quadrant could be colour, the bottom is size, and the right is style (bold, italic etc). Using this, if I wanted large text, I would move to the size quadrant and rotate clockwise to select a large size. Here you are taking advantage of the hierarchy and ranking such a widget could provide.
The font example is really really lame, but the GUI they present has a lot of potential. You can also divide the wheel into an arbitrary number of sections. There are tons of application domains that would work for a component like this. I just hope that 8pen isn't hell-bent only on giving us a keyboard.
I think it's the difference between learning curve vs. optimal speed.
In your Font example (which I think demonstrates your point great), it's easier to jump in to, but your "top speed" is a bit lower than if they grouped it like 8pen chose to.
Not just a bit, but a lot, IMO! Some people are comparing this to the dvorak vs. qwerty keyboard layouts.
A more apt comparison is to a keyboard laid out in alphabetical order vs. a qwerty keyboard. The adoption of qwerty keypads on touch screens is 100% based on familiarity and no ergonomic consideration in letter positions at all.
We're talking potentially major improvements in speed, assuming they have the science of letter positioning correct.
(Please let them have it correct, or we'll have the same dvorak vs qwerty nonsense all over again. Learn from history, people ;))
I guess when I see their product, I don't see a keyboard. I see a cool widget, and I don't mean that in a demeaning way at all, it seems really interesting. In a perfect world, it would be nice to generalize the widget out so it can have multiple applications.
That's not to say the value of the product doesn't lie in the keyboard research they did as well. Anyone can design a cool widget, but the idea to recognize a very innovative solution to mobile typing and perfecting it (figure-8 drawing is a great touch, assuming it works well on many phrases) is 8pen's strongest point. I just hope there is enough flexibility after the patent is issued so I or someone else can write a generalized form of this widget.
I think one fact needs to be understood that the method uses thumb (of a single finger) for all the typing..
I just did a small experiment.. I took the last text I sent from my phone (about 300 characters) and tried pseudo-typing it according to the keyboard the video explains.. and man, my thumb hurts! I would really prefer if there was a way to input text with multiple fingers without slowing down.
wonder how tiring it would be to type everything with this method.
the way its implemented, thumb is only convenient to type using the keyboard.. and if you're right handed, its extremely difficult to master it with left thumb..
Its like learning to write with both hands if you want to switch thumb!
I used this for input when I had an iPhone and it worked surprisingly well. There's also a similar android input method but not quite as well executed as this one.
I don't understand why people are comparing it to Swype- just because both involving swiping instead of taps? In spite of this superficial similarity, don't you see they're solving radically different problems?
This system eliminates inaccuracy and allows true no look typing! Think about what that means for a moment.
Apple execs, if you're reading this, please license this technology and iphone it.
While I am not sure how effective/good 8pen solution is, the problem space they are working in ("touch screen typing"), is bound to produce a winning solution in next few of years. Interesting times.
Like the concept, reminds me of QuickWrite. However, I am skeptical of claims that text could be entered without viewing. At a minimum it seems that viewing would be required to target the center circle. Maybe this could be solved by "re-centering" the entire input space based on initial touch.
The input surface could be marked so that the finger feels what is the home center and where the quadrant boundaries are. The marking could be done with ridges or indentations on the surface, or by vibrations or even sound.
Sure thing. That is another way to hone in on the center. Some sort of tactile gravity well implementation. I think that if 5-ways on candy bar phones were actually joysticks instead of 5 buttons this may work better. And it could definitely work on devices with mini-trackballs.
When they were explaining how to form letters, the movement reminded me of dialing numbers on a rotary phone, except that the possible combinations were expanded because counterclockwise "dialing" is also allowed.
I can't wait to try this out :) It looks like a very cool concept, and while watching the video, I doubted it at first. But I think with practice, that could prove to be the best mobile keyboard ever.
I love keyboards when I'm using my laptop, a desktop, or a blackberry, but I've never gotten the hang of of the virtual keyboards on either iPhones or Android.
the concept is interesting but i doubt it'll catch on. for reasons similar to why dvorak didn't.
using this will require a learning curve, which will mean substantially slower, more frustrating use before your typing becomes faster. and for most of the masses, immediate expediency is preferred. most people will likely switch back to normal keyboards before they actually successfully learn to type on this.
as a side note, i love swype. its similar to this, and doesn't have the same "relearning how to type" issue.
Dvorak didn't catch on? I haven't seen an operating system in years that didn't have Dvorak available as an optional keyboard layout, and I know several people (self included) that use it.
It's not the default, sure, but there's still a critical mass of people using it. There's value in that, much like there's value in a good keyboard-alternative for touchscreens.
Dvorak caught on like Esperanto "caught on". If you're a user yourself you probably know other users, but otherwise you can go your whole life and not encounter it. I doubt more than a few percent of people would even know what Dvorak means, let alone use it.
The only reason there's OS support is that another key binding is trivial. How many computers come with a Dvorak keyboard as an option? That's much less trivial and would be a sign of critical mass.
Google pretty much agrees, there's a paltry 144k results for Dvorak:
How many computers come with Linux as an option? (And how many still would if installing it wasn't significantly cheaper for manufacturers?)
As alternatives go, it's pretty established.
Also, while I know where you're coming from, Esperanto isn't a great example - Dvorak is just as expressive as Qwerty, it just doesn't bend your fingers up as much.
I agree that Esperanto isn't an great example, as it gives more than 40 millions of results on google (7 millions of them if you search only for pages writen in esperanto), much more than Dvorak.
And you imply that esperanto is less expressive than some other language? What language would be that? And from where you got this impression/finding? Or you just afirm it like it was truth having never researched on the subject?
I'm sure that languages based on the roman alphabet are much less expressive than the chinese or japanese, because it is so simple, small and somewhat artificial, isn't? IS NOT!
> How many computers come with Linux as an option? (And how many still would if installing it wasn't significantly cheaper for manufacturers?)
A lot more than come with Dvorak. It also isn't typically cheaper with the way Microsoft charges OEMs. If you're exclusively MS you pay less per copy. There's no QWERTY pusher out there putting the squeeze on Dvorak.
Just a reminder for people wanting to try Dvorak. If you have some spare money on the side, trying Dvorak with a Typematrix keyboard is really worth it. I did the switch, it was painful for 2 months, but now (one year later I think) it is so natural that I wonder how I was able to type on a qwerty standard keyboard before. I will buy another 2030 as spare part, just in case of Typematrix is going broke.
Keyboard choice is a whole other conversation. :) By all means, get a nice keyboard, but you don't need a new keyboard (or labels) just for Dvorak.
I didn't have a hard time finding a keyboard I liked, but it took me a while to find a mouse - I eventually settled on the "Kensington Expert Mouse" trackball (http://www.amazon.com/Kensington-Expert-Optical-Trackball-64...). There aren't a whole lot of left-handed trackballs, and none I've seen have mouse wheels. This one is ambidextrous, the ring around the ball is also a wheel, and it's got four buttons.
This reminds me of Palm's Graffiti system used on the Palm Pilot range of PDAs [ http://en.wikipedia.org/wiki/Graffiti_(Palm_OS) ]. That system too used simple gestures that mimicked hand-writing.
Rather than training the system to learn how to decipher the user's input, it required the user to learn some basic gestures. The input speed may not have been as fast as the predictive touchscreen keyboards on iphone/android phones, but it was a pleasure to use and accuracy was second to none.
I think an important part of what made Graffiti so effective was the fact that you had access to a complete range of alphanumeric and punctuation characters without any need to switch modes. There were actually quite a number of onboard programming tools for Palm devices, so I could easily write little scripts in BASIC, Forth or OnboardC to pass the time or quickly try out an idea. It made the device feel more like a computer and less like a toy.
Most of the touch-based keyboards I've seen on modern smartphones perform acceptably when you're composing a message made out of English words, but you can just about forget coding on one. Even Lua, which is well-known for having a sparse syntax mainly composed of alphabetical keywords, is torture on an iPod Touch.
I remember that. I could actually graffiti faster than my handspring visor deluxe could keep up. The two downsides there are:
1) Needs 2 hands; I couldn't imaging using graffiti with my thumb while holding the device. I type on my phone with my thumb all the time
2) It infringes on Xerox's patents
Swype seems a lot more likely to take off than this. Anyone else think that the finger in their video seemed to move at superhuman speeds? Seems like a lot of finger motion for each character.
This looks to have a lot of potential, IF it can be learned easily enough. This is the first mobile keyboard I've seen that I think could have the possibility of blind typing, but we'll have to wait and see if it is practical.
IMHO, the killer technology for smartphone keyboards would be touch screens that can get textured programmatically. For example, whenever the OS displays the keyboard, the screen would instantly create a slight bevel over each keys. I guess we're a few years away from seeing this though.
This looks pretty neat! I worry that the repeated circles that seem to be involved will lead to some joint stress issues in the thumb that's doing it, though. I'll find out tomorrow...
I feel like this has a potential for use without looking that Swype doesn't; it only relies on hitting one big target rather than precisely and fluidly hitting a bunch of small ones.
Noted mathematician Ken Perlin has done quite a bit of tinkering in this space on similar systems. In the right column, in the second row, check out the "pen input" experiments http://cs.nyu.edu/~perlin/
I would very much like to see an example of a human using the input. No matter how skilled they are - in fact, the more they practiced, the better. I have trouble seeing this as as-quick as a semi-intelligent QWERTY pad that can correct most errors by proximity.
Sure, loops are fast, but are they as fast as a move and tap? And how about ergonomics - is it more tiring, given that the distance travelled is huge compared to many others?
Totally-blind typing: awesome. Huge advantage there where accuracy is important (Graffiti ain't bad, but small errors can magnify as you may cause a command gesture if you aren't watching). I'd call it a definitive "win" there. But how about other uses?
This is measurable- the minimum diameter of a circle passing through the four quadrants is the maximum distance the thumb can ever have to travel. Allowing for some safe radius from the central dot, you would end up with a shorter travelled distance for many letters, even in the worst case. Add to that the fact that qwerty is not optimized at all for the positioning of the thumb- so G is a farther distance away than Q for example- and we're talking major benefits.
Oh, I dunno. Remember that repeat letters with QWERTY are distance-less, while with 8pen they always cost the same (just look at how to type "pizza", as they have it at the end. Even assuming best-case, that's hugely more costly with 8pen).
There are also few advantages / disadvantages with letter combinations, and I'd bet most QWERTY words average close to 1/2 the keyboard-size travel between letters or less. That's slightly less than 1/2 the narrow width of the screen, in straight lines. A single, near-center gesture seems it would exceed that pretty quickly, especially as people are likely to be much sloppier than staying near the center.
This is a cool idea, but if I consider the speed at which I can touch type versus the speed of my handwriting, this may turn out too slow in the long run.
It's too bad the video didn't show an actual person using this system, that should have given a better impression. Also, Fitt's Law may become relevant whenever you try to go back to the circle in the middle without crossing one more line accidentally . The Swype guys seem to have figured that out quite well by predicting which word you're likely to have meant - I wonder if there's anything like this included in the8pen, too?
(P.S.: Justin Timberlake says: "Drop the article in the URL")
It's really just a matter of implementation to allow "soft" detection of gestures and make the estimate using the probability distribution of words and letters in the dictionary like swype. It seems that there are hard boundaries in the visual system, but I guess it's just gesture detection at heart, which makes it pretty robust and fast in practice.
I've been actually prototyping a pretty much similar system, but I'm not sure if it's a worthwhile venture. I guess it should be clearly better to have any competitive advantage.
Its out now, and I was going to try it like a lot of other people have said, but there is no way to try it for free. The cost isn't very high ($1.58 right now) but with Swype already I don't see a reason to pay money just to try it.
I just picked this up. I have very large hands so I'm always looking for an alternative to a standard on screen keyboard. This works amazingly well. I'm already becoming proficient with it, and it is nice to build up a natural rhythm once you get going. I am already 100% convinced and switching completely over. I was slightly annoyed at first by the lack of a free trial, but I figured this was interesting enough technology that whether I liked it or not, I was willing to spend $1.58 to find out, and it was definitely worth it for me. Highly recommended, that said, its not for everyone.
The video takes too long to get to the point. A better approach might be to sell the performance upfront very quickly, by having a video [ autoplay ? ] of a real user entering real text ie. 8pen on iPhone actual footage. Only then is it obviously worth the viewers time to go in more depth to see the explanation.
Im presold on the idea that there must be a better way - swype is one well known improvement. I dont need to be sold the problem, I need to see the solution working in a real scenario before exploring further.
This reminds me an awful lot of a system presented [1] in lectures on accessibility research a couple years back that was/is an alternative gestural input method for people with motor impairments. It's cool to see similar concepts cross over into the mainstream...well, sort of mainstream, anyway.
Dumb question here, but why don't we simply use handwriting recognition?
Have a space at the bottom that is big enough to write two characters, so the system can know when you've started the second character, you are done with the first.
I could write with that super fast. And HW recognition is probably better, for most people, than their mobile phone typing accuracy.
If this is properly optimized, it should require much less finger movement than handwriting. And it can easily get even casual users to higher accuracies than even full size keyboards- the recognition is much more accurate because of the four quadrant, set, layout.
What I love about your comment is that it has an actual speed.
The answer on this is simple. Take random groups people. Give group swype, give another group this. Train them for 10 minutes and test them. Report speed and error rate, bindfolded and normal. The train them for another 20 minutes, retest, then 30 minutes a day for 5 days and retest.
Tried this. The concept seems quite nice. Problem I'm having is that my thumb sweats when pressed to the touchscreen and creates friction which makes doing the continuous gestures unpleasant. Rubbing talcum powder into my hands seems to help with this.
Still, I'll see if I can learn to write a bit more fluidly with this.
Based on all the comments here (and the comments left on swype related threads) I have to wonder if I'm the only person who is genuinely happy with the iPhone keyboard? I can type extremely fast and extremely accurately on it. I honestly don't think either Swype or 8pen would improve my wpm.
I'm also very jealous of options like Swype and now 8pen, but I'm fairly satisfied with the iPhone keyboard in terms of inputting individual letters. What I don't understand is how Apple thinks that their current auto-correction is acceptable.
Typing Japanese directly on a phone's numpad should be a huge pain, but the phones have for years featured very intelligent prediction that makes it so you can type much faster than English even though there are five letters per key (as opposed to three in English). The phones come with standard "dictionaries" and also learn your own patterns, so you rarely ever have to actually type more than one letter of a word. Based on the rest of the sentence, there are really only a few options of what could come next and the phone can figure it out from one letter.
Whenever I'm typing on the iPhone and end up with crap like, "Okay, then ill send you the file," I can't understand how something like this is acceptable in 2010. There is absolutely no reason the iPhone shouldn't know what I'm trying to type; even a simple grammar checker could solve this problem. A more sophisticated prediction algorithm could drastically speed up typing on the iPhone without changing the interface at all. If nothing else, at least stop making me type apostrophes.
I really hope Apple sees the vulerability in their aging keyboard and buys someone like 8pen or Swype so they don't fall way behind.
On my iPhone 4 it correctly adds the apostrophe after typing "Okay, then ill". As far as I can tell in all situations the phone considers "ill" to mean "i'll". If you really meant "ill" you just have to tap the suggestion to tell it no.
I can't see Apple going the route of Swype or 8pen, too confusing. Apple likes simplicity (how many decades before they succumbed and released a two button mouse?)
It's hard for me to say intuitively regarding speed and the Swype, but I can see that Swype would probably increase accuracy by some statistically significant amount.
This device, based on my intuition, would probably offer significant increases in wpm, and definitely increase accuracy significantly.
It's not a question of being genuinely happy or not- if something better comes along, than obviously it should be adopted.
The video never shows it in use. That's because even after you learn it, it's going to be much slower than typing. "Just like handwriting"... guess what; handwriting sucks http://lumma.org/microwave/#2007.07.12.2
1) The keyboard (layout) was not invented for the computer, but for the typewriter;
2) Their whole premise seems to be based on 'gestures that mimic handwriting', well why don't they just have gestures that are like handwritten characters?
Gestures that are like handwritten characters are harder to deal with on the software side and slower to use on the human side. Handwriting recognition has been around for a long time. I don't know how this will compare speed-wise, but it's definitely simpler.
I still think that traditional shorthand with error correction is the fastest way to go without keyboard. Any system can be used, e.g. Gregg, Pittman, Merlin, or a new system can be invented. Of course the learning curve is substantial, though.
Having only looked at the pictures, it reminds me of GNOME's Dasher, which I have never really had trouble with on the few occasions I've played with it. So that's good because on-screen keyboards are pretty sucky at the moment.
Currently on an HTC Hero with Android 2.2 The problem with the Hero is that the keys are way too small. I have normal sized fingers (for a male) so I reckon I'm not the only one with this problem!
Also, if it's not ready to download until tomorrow, why not just announce it tomorrow? I would try it right now, but I know I'll have forgotten about it by tomorrow.
I use swype now and am extremely happy with it. I can't imagine this would be any better than it. With swype, I can enter n characters with a path of at most n vertices, with this thing, it seems I need to create n differently-sized arcs. Swype also has zero learning curve, assuming you know qwerty, because you can use it as a touch keyboard by default.
It's not clear if the finger needs to be removed between letters (but I doubt it does). This would worry me.
