Typing drills for new users

[fingerworks]

Ultimate success with Zero-Force surface typing requires that you:


[list=1]
Get accustomed to resting all 10 fingers on the surface simultaneously (and regularly). So long as all 5 fingers in a hand drop simultaneously, they will NOT cause any spurious keypresses.
Learn NOT to rest just 2-3 fingers at a time during typing, since the typing recognizer has no choice but to interpret these as (unwanted) click chords or keypresses.
Minimize excursions from home row.
Quickly return to home row after long reaches.
[/list=1]

I recommend the following two drills to quickly improve these skills:
[list=1]
After typing each word, pause and drop all ten fingers on the surface, feeling for the home row dimples. Then hover and type the next word. Repeat this, resting between each word, for several minutes. Once you can rest and re-center 10 fingers subconsciously, you'll only need to do it at the end of a thought or phrase.
Imagine your hand is on a spring, so that each time you have to reach for a peripheral key far from home row, your hand quickly and accurately snaps back to home. It takes a few weeks for your brain to master these accurate reaches and returns for each peripheral key, but this is the skill that ultimately compensates for the lack of key-edge tactile feedback. Actually, maybe I should say whole 'arm' on a spring, because you should always keep your wrist straight 'N steady (sliding your whole arm) during long reaches.
[/list=1]

One can imagine an ideal surface typing layout would have just three rows of keys, dispensing with long reaches altogether. Most punctuation, number, and function keys would be 'folded' into these rows and activated with opposite-hand modifier chords. Such a radical layout may not be practical until surface covers include a dynamic display that changes the visible key symbols as different modifier chords are pressed.

However, if you're willing to memorize embedded keypads without their symbols visible on the cover, you may want to go ahead and move those peripheral punctuation, number, or function keys that you use most often into the three main rows. This can be done via the gesture editor's surface key customization tool.

For instance, programmers who frequently use punctuation will definitely want to enable Enhanced Modifier Chords and the Programmers' Punctuation Pad. Then by holding a thumb+3-finger chord in one hand, the opposite hand can type most punctuation symbols on or around home row. With the MyGesture Editor, you can create a similarly activated number or emoticon pad for the left hand.


Wayne Westerman
FingerWorks CTO
& The First Zero-Force Typer :)

[Pug]

I've been struggling for three weeks with my Dvorak Touchstream LP, and through some insanity my hands have come to figure out where the keys are using the edges of the surface as a reference. This can't be optimal, but it does appear to work.

The 'spring' idea seems curious to me, though, so I'll try giving that a shot.

Thankee, O CTO master.

[piusvii]

I will try this but I'm somewhat skeptical. Long periods of typing are still the main difficulty with my TS. Its wonderful for mouse fns but prolonged typing remains more tedious than std keyboards. One thing I noticed is that the key layout, even on the split TS, has the keys on left board extending up and to the left slightly. On typical split keyboards, the keys are angled slightly towards a center point so that the hands and fingers can maintain a more natural position at the wrists. Any chance of this changing in the future?

[nomaded]

Well, looking at those suggestions about returning the 5 fingers to the home dotted positions, I find that in order to put all my fingers on those raised dots, and have my fingers curl naturally, I need to angle my forearms a bit more than I would normally. I think if you have your arms in this position, the keyboard layout works the same was as if the keys were angled more centrally as on a microsoft natural keyboard. I think it's just a matter of changing one's arm position a bit, to keep the wrist in a natural position. Give that a try. I know I will, to see if it helps my typing accuracy.

[-jeffB]

Any suggestions for classical guitar players and others of us with longer fingernails? I'm having some difficulty making keys on the bottom row of the right hand register. I'm trying to move my whole hand down, but it seems much easier and more natural to just curl my fingers under, and that's where I get in trouble.

(By the way, the built-in autocorrect appears to be doing an excellent job!)

[piusvii]

I haven't found any alternative to cutting my fingernails often. It has made fingernail grooming much more consistent but I guess it would be a problem for guitar players or those who really like oranges. :)

[Dvorak!=Qwerty]

I've had TS LP for about 3 weeks. Still on the learning curve and improving, but here is a question i wondered. If the letter portions of the keys were slightly raised (few milimeters perhaps) it seems to me like there would be sufficient tactile feeling for feed back to reduce/eliminate hand drift. I also think a slightly 'bumpy' surface wouldn't impede mousing/gesturing a whole lot (since downward force is so small, the fingers would glide over valleys). Quite possibly this could make sensor plate design more difficult.

I find it hard to think that it wasn't considered. I wonder what the trade offs were. Any insights?

[fingerworks]

Many people ask this. The problem is that any noticeable edge contouring throws the position interpolation by the sensor array completely out of whack. (i.e., you would be sliding your fingers along, and the mouse pointer would hop or jiggle whenever you passed over an edge). Even the home row 'braille dots' cause a slight, but usually not noticeable, deviation.

It may be possible to avoid this problem with a MUCH higher sensor density, but that is not economically feasible yet. Maybe when the surfaces are all built with plastic transistors :)

[-jeffB]

On a related note, are there any white papers or academic pubs that talk about the particular technologies used in the TouchStream products?

[John Meacham]

Instead of raised edges on the letters, a slightly different tactile feel. like a matte vs glossy finish on the plastic. or perhaps a thin aluminum foil type layer which would be just as slippery but cool to the touch so you can tell when you are drifting, not sure if any of this is necisarry though, I was having the same thoughts when i first started using it, but now I seem to be drifting much less.

Speaking of drifting, would it be possible to asssign a keycode for the touchstream to output when it detects you are drifting off the home row? using 'xbindkeys' and 'osd_cat' i could very simply make a big message Whoa! appear overlaid on the screen when the machine recieves that keycode. I think this would be quite cool and help preemptivly avoid drifting. as it is, it is hard to tell until the wrong letters start coming out, something that is not always obvious when you are doing things other than typing text such as working in vi command mode. It should be easy to implement on the touchstream side. (I think)

People could also map the keycode to 'beep' if they feel like making their keyboard noisy again :)

[Aloft]

What would be even neater, if you are going to detect drift, would be to actually compensate for it. That is, move the soft location of the keys as the hand drifts, and when any other action, such as a mouse click event, happens it automatically resets. It could also pretty easily tell when the user was resting and feeling for the dots and reset at that time too. I think the idea that the location of the keys is not fixed could potentially be one of the strongest points of this technology in the future. Imagine keyboards that automatically compensate to the user's hands over time. For instance, I have short fingers, so I consistently come up short on my Ys and Qs. It could detect a large number of h-backspace-y sequences and slowly bring the distance of the Y down to where I can comfortably reach it. That is just one example. When the visual display becomes dynamic as well, then we'll start seeing Star Trek like interface advances. Neat stuff. I firmly believe this tech is the way of the future, and will laugh nostalgically at how crude these early designs were. :)

[John Meacham]

happily enough, the touchstream already does this :) at least that is what I am led to believe. but it can't compensate for just absurd drift, which is why I wanted a warning just before hand.

[JerryKnight]

"3. Minimize excursions from home row."

This means learn dvorak ;-) Seriously, the layout has the best use of the home row that I have ever seen, and it is far better than Qwerty in this respect. I have never tried Qwerty on the LP, but I can only imagine that it is much harder.

Granted, learning both a new layout and the ZF typing would be a serious blow to the sanity for a while. I made it a point to learn Dvorak before getting the LP. ZF was hard enough to learn without the brain cramping chore of de-qwerty-ing myself.

It has to be acknowledged that qwerty makes inadequate use of the home row, which is now an issue since the only non-visual "anchor" on the LP is the dots on the home row. Not everyone can or should go through the headache of learning another layout, but it would definitely help on the accuracy.