Slashdot Mirror
Finding an Optimal Keyboard Layout For Swype
New submitter Analog24 writes: The QWERTY keyboard was not designed with modern touchscreen usage in mind, especially when it comes to swype texting. A recent study attempted to optimize the standard keyboard layout to minimize the number of swype errors. The result was a new layout that reduces the rate of swipe interpretation mistakes by 50.1% compared to the QWERTY keyboard.
Since the QWERTY layout was designed to SLOW down typists of the day, so the mechanical typewriters of the day wouldn't get their keys stuck or jammed together from typing too fast before the previous letter's hammer fell back.
-- You are in a maze of little, twisty passages, all different... --
I can't type and I blame QWERTY
Back when I had a Palm Pilot I found that the Graffiti entry method was very effective, much more than trying to press tiny on-screen 'keys' with a fingertip almost ten times bigger.
Instead of a stylus, how about we make the area of the on-screen keyboard instead act as a finger-pad when the phone is held in-portrait? I think it'd be big enough to reproduce Graffiti strokes with one's finger so that a stylus wouldn't be necessary...
Do not look into laser with remaining eye.
Tried many times in the past, failed. The amount of time and effort to reach proficiency on a non-standard layout, combined with the presence of standard layouts in business and consumer products, means you are fighting the Beta vs. VHS war. It's right up there with designing new pet programming languages that are essentially syntactic sugar, offering no additional computability.
I couldn't even adapt to swyping.
...gis sdrawkcab (usually not responding to ACs; don't bother posting as AC)
The effectiveness of swyping depends quite a bit on the language. I find that it works fairly well in English, but not nearly as well in some other languages, where there may be many more ambiguous words. This also means that the optimal keyboard layout would depend on the language, which would be horrible to use for those of us who use two or more different languages.
TL;DR: Classic keyboard layout optimization places commonly used keys close together to reduce finger motion, but swipe layout optimization spreads them out in order to improve recognition.
SwiftKey is extremely fast, especially with Flow get has totally revolutionised smartphone usage for me.
reduces the rate of swipe interpretation mistakes by 50.1% compared to the QWERY keyboard.
I think you accidentally a letter.
systemd is Roko's Basilisk.
Swyper, no swyping.
Zero chance of retraining people's brains. Any keyboard layout requires muscle memory. Doesn't matter which one you pick, you still have to use it a long time and develop the muscle memory.
This is where UI designers go horribly wrong. They don't realize that people have been using View / Source in Firefox, or File / Save As... in GIMP, etc for LONGER THAN THESE PEOPLE HAVE BEEN OUT OF COLLEGE and it's next to impossible to override years of muscle memory. Even if you know in your brain that View / Source has been moved, or that it's File / Export... now, you can't stop the muscle memory. That's why so many people hate UI designers who shuffle things around for no reason. It doesn't matter where View / Source is. What matters is that it's always where it is supposed to be.
Keyboards are becoming obsolete. You can text without one on a smartphone. About the only people who will use them in 20 years will be coders, because of the different conventions for #defines, variables, etc.
Remember Mr. Scott trying to use a computer by talking to it, and then picking up the mouse and speaking into it?
"Transparent" is a shit show that trades on every stereotype going. A man in drag is NOT a transsexual.
You're right, people do get lazier after they adjust to using a new input mechanism and their error rate goes up. The same thing would likely happen if people used the keyboard we presented but there would be fewer errors than you get on a QWERTY keyboard. The main point of the analysis was to reduce the total number of possible errors.
The purpose of Swype is to speed up typing on a touchscreen in an easy to use and intuitive way.
How do you think completely changing the layout that people have been learning for the past 140 years will be received? If swype didn't fit with existing keyboards no one would use it.
https://play.google.com/store/...
I used that app back in 2012, and it wasn't just-out-of-beta then, either. When I used it, it effectively solved all of its intended problems: common letter combinations were in close proximity, and there were fewer "word collisions" than with Swype ("or" vs "our" immediately coming to mind).
The problem I found was the fact that while Flow is more accurate, it took me significantly longer to type out a message because I wasn't used to the layout. I've spent over 20 years on a QWERTY keyboard, and even though Flow is more efficient, QWERTY is basically like breathing for me, so any statistically-better keyboard would still be slower because I'd be re-learning to type all over again. Even with errors, I'm personally still a Swype person at heart. I've even tried Swiftkey and Go, but even there what holds me back is knowing where the punctuation keys are. The only alternate keyboard I'll use is "Hacker's Keyboard" when I find myself in an SSH session, but other than that, Swype got me in early in the WinMo 6.5 days, and has never let go.
Flow is another keyboard designed along similar lines, though it's optimized for slightly different criteria: fast text entry rather than low error rate. It takes a while to get used to, but you really can type a lot faster with it.
"I'm too busy to research this and form an educated opinion, but I do have time to tell everyone my uninformed opinion."
The important thing with novel keyboard layouts is that people need to be able to feel confident that they will stick around. I'm not going to want to learn to use a new layout if I don't think it will still be a viable option in three years' time. It takes a long time to build up proficiency with a new layout.
Personally, I think the optimum solution would be some type of hardware-based, single-handed chording keyboard. Unfortunately, that suffers the same problem to an even greater degree, because it takes even longer to learn how to use such an input device.
The work of finding fast input methods was done more than a century ago and it resulted in a wealth of different Shorthand systems, see: http://en.wikipedia.org/wiki/S... systems. I'm still waiting for someone to write an input method using one of these systems. This should be much faster than even an optimal swype based system. The downside is the learning curve, which is probably very long.
Nobody here has used hipjot (formerly NIN I believe) for iOS?
You can type faster on this than you can with a physical keyboard, easily.
https://youtu.be/E7ktz8pHbYc
Alas, simply use the American English keyboard. Other parts of the world can use their local toys if they want to be different.
A random walk with hops being shortened over time is called "simulated annealing". It's an alternative to genetic algorithms and tends to be easier to use for problems with solutions that can't be chopped up and put together in a coherent format. For instance, keyboard layouts, which require each key to be present exactly once.
QWERY keyboard layou is the bes, you insensiive clod!
---- MISSING MISCELLANEOUS DATA SEGMENT --- [sigdash] trolololol