Here’s a fascinating video demo of Dasher, a single finger interface that allows you to enter text without a keyboard. Single-finger input without a keyboard? Wait, that’s my iPhone.
Keyboards are inefficient for two reasons: they do not exploit the redundancy in normal … all » language; and they waste the fine analogue capabilities of the user’s motor system (fingers and eyes, for example). I describe a system intended to rectify both these inefficiencies. Dasher is a text-entry system in which a language model plays an integral role, and it’s driven by continuous gestures. Users can achieve single-finger writing speeds of 35 words per minute and hands-free writing speeds of 25 words per minute. Dasher is free software, and it works in all languages, and on many platforms. Dasher is part of Debian, and there’s even a little java version for your web-browser.
38 thoughts on “Single-finger text input”
Comments are closed.
ADVERTISEMENT
Join Make: Community Today
It would be faster, easier, and vastly less complex computationally to learn Morse code. Text entry speeds in Morse, the original one- (or two-)fingered mode, can easily exceed 35wpm with moderate practice, and it doesn’t require a screen or even vision.
Receiving Morse code is a different story – much harder to learn than sending, and it takes extensive practice to build up to even 20wpm. If you’re using it for text entry, though, you wouldn’t have to be able to receive it at all.
wow this works quite well, even for writing non-dictionary words.
Major downside is the size of the entry box required. I can’t see this working on a cell phone, unless its reasonably sized and the whole screen is used.
alandove: you overestimate the average ratio of human effort to computational effort :)
I have a feeling some folks might get motion sickness from using Dasher.
Well, this is quite interesting, but, I agree it would be difficult to use on a (even not so) small cell phone…
I think alandove is right about morse code, but I think it will create new problems (learning and RSI, for example). And what with languages which does not use letters or use non-latin letters (arab, japanese, german, russian, etc.)
About languages which does not use letters, I’m quite curious about how Dasher will deal with such languages.
Let speak about Japanese, for example, will you display a list of letters to compose kana or directly a list of kana !?… I think choosing the kanji is quite a different problem since you still can go through a two-step validation (one for letters/kana then one for kanji) without too much difficulty ??
I remember something more or less identical about 5 years ago called “runline” (can’t find any links). I’ve played around with it on the Zaurus PDA and it works on that screen size (pretty close to the IPhone). Biggest potential I see is for quadriplegics and people with locked-in syndrome to be able to interface with computers more easily.
This item for double-post, so I deleted the other one. Here’s a comment from that one:
“I played with Dasher several years ago (quite a few years ago). I was wondering the other day why I had never seen it ported to any modern devices, but I couldn’t remember what it was called.
I experimented with it on an early generation Pocket PC and it was fairly fast and accurate (and a bit hypnotic). Seems a perfect match for accelerometer control.”
I’ve played with dasher. The trouble with it is that you need to read nearly randomly positioned moving letters and *find* the one you want. I can definitely type faster on an iPhone.
On any keyboard-like input system the letters don’t move so you can remember where they are. Imagine trying to touchtype if the letters moved!
Still, it’s a neat idea.
Gareth: Judging by your comments and what I’ve heard from Dasher people you may be pleasantly surprised by some of the things they’re looking at for the future. I’m not going to scoop their future announcements by saying any more right now tho ;-)
I think people are treating this as a complete technology when they say things like “I can type quicker on my iPhone”. This is a demonstration of a possible implementation – there would undoubtably be tidying up of it, research into how people would use it, it would be refined and improved… etc.
Also, the other mistake people are making when they say “I can type faster than this on my iPhone” is that they are forgetting this method could be consistent on many different devices – you learn it once and the same method can be applied to your iPhone, your PC, information kiosks, etc.
there’s a fine line between clever and stupid – spinal tap
People who are nearly fully paralyzed and retain only the use of one arm or even only their eye movements could make great use of this.
The structure is interesting. Definitely more efficient than several current systems. But, the static notion of a “library of all possible books” implies that language doesn’t evolve or develop.
Luke: I think it is more useful to think of it as a collection of all possible strings constructed using a given alphabet. The speaker’s phrase “a library of all possible books” sounds poetic, but that’s what he really means I think.
I think the most “wow” factor for me out of this can be summed with the part: “By navigating, I am writing.” That’s definitely the paradigm-buster right there. The linguistic structure aspects seem no more than an application of T9, but I anticipate that’s not the primary focus. I’m not clear what would be the optimal implementation of this graphics-wise, but this most definitely has future potential, maybe with haptics instead to allow sensing of the sentence topology rather than visually navigating, or something of the sort…
this might or might not have been experimented with, but what if:
*more weighted-suggested letters appeared at a progressively larger scale
*compiled text box was placed inside the nav box, smaller width, and scrolled with the completed text, to allow better peripheral focus and hence efficiency
*as letters are chosen, animated ‘magnetic’ letters zip to text box
*existing predictive word technology embedded to choose words as well (plus space character on end of word) Words suggested at top of the list. eg. H > E > L > choose words ‘Hello, Help, etc’
I’m not sure what some folks are finding so wrong with Dasher. It is not taking away your keyboards. What’s so bad about providing an alternative input method for folks who, perhaps, can’t use a keyboard?
alandove: I’m not so sure learning a new alphabet (essentially what Morse is) would be easier than learning to direct Dasher. One can build sentences as soon as you begin with Dasher, without consulting a lookup table as beginners would likely do with Morse.
Tait: Dasher’s been around since 2002–earlier if you count David Ward’s original research.
What I find most interesting is the similarity to arithmetic coding (that something Ward’s research centered on, I believe). You navigate towards an increasingly precise point along the single-dimensional axis, and the string you put together is what would be encoded were it actual arithmetic coding. It’s a great illustrator in CS courses.
Looks very interesting. I’d switch it up so that it starts guessing what words are next with the ability to go back to the letter navigation.
And the UI is a bit hard to read. I’m assuming that after you’ve used it for awhile you get a sense of where the letters you want are, and navigate there subconsciously.
I’m very interested. Nice one, Make:. More posts like this.
Absurdly complicated and totally useless. While he’s thinking about it, I can type the words manually. I’m not even a good typist. A good typists can do 60 WPM and more. That makes this look as silly as it is.
Would you PLEASE just take the time and effort to get a speech recognition program that works?
Better yet, direct neural input. Where’s the chip in my head that I was promised???