h a l f b a k e r yNot so much a thought experiment as a single neuron misfire.
add, search, annotate, link, view, overview, recent, by name, random
news, help, about, links, report a problem
browse anonymously,
or get an account
and write.
register,
|
|
|
Instead of typing a different key for each letter, one could use an editor with a line of letters (and space other characters) to the right of the text. By using [->] followed by [Enter] the next letter to be added to the text is chosen from that line.
Better than having an alphabetic line, would
be to have the letters arranged in order of frequency of use: etaoinsrhldcumfpgwybvkxjqz. Another possibility is to have the cursor start each time in the middle of this line: qxvygfudhsiaetonrlcmpwbkjz and use both [<-] and [->].
With practice and a fast cursor, one could two-finger type quite quickly.
how to talk when your'e in a coma
http://slate.msn.com/id/2113353/
they could use your help fj [xclamp, Feb 12 2005]
Dasher
http://www.inferenc...y.cam.ac.uk/dasher/
Uses the mouse (or eye tracking software) instea do cursors. [ironfroggy, Feb 15 2005]
Dasher Buttons
http://www.inferenc...ment/Breathing.html
This page talks about the original version of dasher, which is pretty much exactly like the idea described here, using buttons. [ironfroggy, Feb 15 2005]
Please log in.
If you're not logged in,
you can see what this page
looks like, but you will
not be able to add anything.
Annotation:
|
| |
Reminds me of the old "band printers" we used to use when I was learning to program in college. (Boy, I'm dating myself). |
|
| |
Would be great for replacing the tiny keyboards on phones, etc. Add predictive software to the mix and instead of the cursor starting each time in the middle, have it land on the letter with the highest statistical probability of being next. Could probably type fairly quickly. |
|
| |
Also see (link) for a cool alternate "mind controlled" application for this. |
|
| |
Any programming wizards that could whip up a prototype? |
|
| |
"Another possibility is to have the cursor start each time in the middle of this line: zjkbwpfudhsiaetonrlcmpwvxq and use both [<-] and [->]." |
|
| |
To make it easier for beginners to find the letters, you could put letters from the beginning of the alphabet above the middle point.
jkbgfmcdlhiae.tonsrupwyvxqz |
|
| |
It would be interesting to see how the order-by-frequency would change for different languages. |
|
| |
In your last example, I think one of your P's should be a Y. |
|
| |
I was curious about something similar that wouldn't rely on fast reaction time (thinking of a quadriplegic gentleman that I know who uses a mouth stick), so I hacked together an application to test: |
|
| |
Using the left and right keys, a binary search algorithm can permit selection of any of the 26 characters with a maximum of 4 [<-],[->] keystrokes. |
|
| |
Using the 4 arrow keys (down = accept current character, up = go back) and one or more fingers, I can enter text reasonably well without any frequency of use or predictive algorithm. (The frequency of use layout could be easily tested). |
|
| |
It might be good to explain what binary search means...You could say that the arrow keys take you halfway to the end of the menu, or halfway back. |
|
| |
With the binary search interface, it's not obvious to the user exactly where the cursor will land when he presses an arrow. But you could highlight or underline the two letters that would come next. |
|
| |
//With practice and a fast cursor, one could two-finger type quite quickly.// |
|
| |
[robinism] Fixed, thanks. |
|
| |
[r], no need for me to explain "binary search", you just did. Actually, the binary search method does take a second to get used to because of the overshoot, but really it's just a series of directions once it's implemented in this interface: "it's to the left, it's to the right", it's actually fairly easy. It doesn't seem to me that indicating the next letters would add much value, you still have to move that direction. |
|
| |
Just plunking in the alphabet in its normal order, it takes 44 keystrokes to spell "HALFBAKERY". Modifying the frequency of use information to correlate to the earliest hits in the binary search (successive approximation might be a better phrase?) algorithm, it takes 36 keystrokes. (including the "accept" keystroke) "CAT" requires 14 strokes with standard alpha and 8 with the optimized. |
|
| |
I know, it has nothing to do with the idea other than to facilitate text input with a minimal number of keys, and nobody cares, but that's what I found. Probably could have found the same thing with a Google search. |
|
| |
It seems that FJ's method is basically what's used to enter your initials on arcade/video games when you get high score (do they still do that?). I sort of remember even pinball machines doing something similar by using the flipper buttons. It's certainly workable. |
|
| |
"Probably could have found the same thing with a Google search. " |
|
| |
I tried google and got this: |
|
| |
'Your search - "half's discoveries" - did not match any documents.' |
|
| |
'Suggestions:
- Make sure all words are spelled correctly.
- Try different keywords.
- Try more general keywords.
- Check the Halfbakery' |
|
| |
& ... brevity is the soul of wit. |
|
| |
Imagine the time consumed by formulating a simple phrase if the search constraints were like the game 'charades'. Would an intellegent system conclude what you were trying to say, and prompt you leadingly to choose among increasingly more narrow and specific choices until you'd composed exactly the message you wish to relate? |
|
| |
This is like entering your initials on the high score list at the end of a video game. |
|
| |
That was quite slow, actually.
Maybe a two-dimensional grid would be better? Shaped like a plus sign. Vowels up and down the center, consonants left to right. |
|
| |
The one-dimensional menu has the advantage of taking up less space on the screen. |
|
| |
[half], that program you posted is super-duper. By combining FJ's idea (Arrange letters by frequency of use) and your idea (binary search) you get a usable interface. I used this sequence with the binary-optimized setting:
BCGHJDAKFPIQLEVMWNXRTYSOZU |
|
| |
Isn't this basically like entering your name on a video arcade game scoreboard? The idea is pretty baked, and using cursor keys for it is going to be pretty slow. A joystick or pointing device is going to be much more efficient. (see my link) |
|
| |
I’d like it with binary search. I’d also like it with a built-in customizable dictionary to do predictive text similar to the way it’s done in T9, but with this two key interface. That would speed it up significantly. |
|
| |
Actually, I just thought of an even better technology to use with this. It may still be in development though. Will search for link. |
|
| |
The Dasher Project - thanks [ironfroggy]
I tried the demo of this a few years back. It takes about 10 minutes to get used to, and then it’s incredibly easy to use. I remembered it as requiring up/down only, but it’s actually x/y, with x controlling the speed, so it may be hard to operate with two keys. |
|
| |
[Shz], I don't think using cursor keys (as opposed to a pointing device) would be any faster to begin with, so that point about two-keys and the x/y controls of dasher is kind of lost. |
|
| |
Ah, but I’m not suggesting cursor based input for Dasher. I’m suggesting that FJ’s cursor based input device would be faster with word prediction software, not necessarily Dasher’s implementation of it. |
|
| |