ULTIMATE INPUT

[Input Nirvana]

Attending Keycon, besides being fun and THE BEST THING IN THE WORLD, has also re-stimulated dead areas in my cranium.

I keep revolving around these same products, knowing they are drastically imperfect and tarnishing my ideas of INPUT NIRVANA for the last 5 years. 5 YEARS! I can do better than be in a stalemate for 5 YEARS over something so simple.

Look:
-Datahand
-Alphagrip
-Split Kinesis
----------------
-RollerMouse
-Apple/Logitech Trackpads
-Integrated mousing
----------------
-Software/firmware

As far as I know input is, can or will be either via mental, auditory, or physical.
Mental is problematic, I'll let someone else solve that design issue. Wires coming out of head or a massive Bluetooth chip glued to skull?
Auditory is acceptable in only certain circumstances such as you alone in a room and not listening to music.
Physical is our domain, for better or worse (obviously worse since we spend so much time tinkering with it).

My latest incarnation was to take the Split Kinesis key wells only (no casing) and make them hand-mounted. Not chair mounted. Sooooo…….

This is IT:
Datahand-like 5 direction fingertip actuators mounted into a half-glove skeleton (literally a wire-frame) that is semi-firm for 4/8 fingers. Thumbs may have more/less. If there is chording, it can be on one hand, no chording it will be on two hands.
Can be slipped on/off without using the other hand, it will be a slip on from the side so it doesn't fall off, very little weight. Palm mounted, so no wrist movement.
Electronics could be remote.
Fingertip movement will be LESS THAN DATAHAND so the not-so-awesome side-to-side MOVEMENT becomes side-to-side PRESSURE….very different.
Mousing intergrated, like Datahand.
May not be finger tip, may be finger actuated, you know what I mean?

Possible drawbacks:
May not be the fastest for 100WPM.
Partial glove/mechanism on hands, different that placing hands on something or holding something.
Durability.
Portability.

This would eliminate any RSI, misc. health related problems. We're talking about greatly minimized movement for fingers and no wrist positioning issues. Missing/fused digits are a problem. I have 80-85% of this in my head right now. Can't be much of a stretch to fill in the 15-20% of incomplete/difficult to implement balance.

Technology exists for this. Skill exists for this. Manufacturing exists for this. Design exists for this. Everything needed exists and I saw it all Keycon West 2014.

Does anyone follow this train of thought? Do I need to give more detail?

[kurplop]

Heeeeee's back!

[Zekromtor]

[TotalChaos]

This would eliminate any RSI, misc. health related problems.

Just to clarify: you will never eliminate RSI so long as you allow impacts every time someone presses a key.  You MUST make a switch that makes impacts impossible, for instance by using a spring or 2.  Jarring impacts will always cause health problems in all living humans.

[Input Nirvana]

Reducing RSI.
Hopefully reducing to the absolute minimum

My thought is that the switches are activated by the movement, which is as soft or softer than a Datahand   And to reduce the amount of movement as well.

[Oobly]

How about strain gauges for the sideways pressure detection, they can be really cheap and the mechanism simple. For one handed use you could possibly even use combinations of press and sideways for 8 different functions per finger. So unpressed + direction is one function, pressed + direction is another. That's 48 key functions before you even get to the thumb. I think the thumb could use only 4 directions and assign them to functions that can sometimes be used in combinations, such as Ctrl and Alt, Fn layer and Shift, etc. See my keyboard build thread for thumb cluster functions. You then press diagonally for Ctrl + Alt or Fn + Shift.

One possible issue I see with the design is that you actually do need to move your fingers now and then, or just hand positions. Staying in one position with repetetive pressure in different directions will also end up causing strain on certain muscles and tendons.

If possible, perhaps a glove with magnetorheological or electrorheological fluid pistons built into it that when no current is applied allows the fingers to move as normal, but when switched on, the pistons resist movement strongly and can be used for pressure detection? Just throwing ideas out there. So you can have almost normal movement when the device is off, but switch it on and it becomes stiff and senses your inputs. Possibly with a thumb press to turn it on and off? Then you can type a bit, turn it off, move your hand, turn it on and continue typing. Much like you would rest your hand on a palm rest between active typing bursts.

That's what I'd like to use

[Input Nirvana]

I'm convinced the final stage in ergo while still using fingers is a semi-ridgid skeleton glove using Datahand motions. This would reduce the stationary (un-ergo) palm Datahand palm rests. The technology is there. I'm wondering if I can have time to do something with it.

[jacobolus]

The problem with any kind of sideways motion is that the ring and middle fingers and to some extent the pinky don’t really have joints which provide very strong or independent side-to-side motion, so when you try to put pressure with one finger you actually end up moving multiple fingers or even the whole hand. Speaking only for myself I find these motions awkward, slow, and imprecise.

For something mounted in a glove or whatever (I'm assuming the goal is to be portable enough to walk around with this thing?), I think it's probably sufficient to just have 2-3 buttons per finger (perhaps more for index finger and thumb if you really want), and then use chords. For instance, if each finger has 2 switches under it, you can press one, the other, or both, so that's 3 movements per finger.

If for the four fingers of one hand, we allow chording, such that we can either do nothing or we can do one of those three movements (and use the same movement across all active fingers), then that already leaves us with 15 combinations of fingers * 3 movements = 45 possible inputs just with the four fingers on one hand.

Now if we have say 3-4 thumb "modifier keys" that gets us up into the hundreds of actions just with one hand.

[Zekromtor]

In order for this device to qualify as ultimate it can't really use chords in order to stretch its usable number of inputs. That is a sacrifice in efficiency. Sadly, I think you're right about the finger motion though.

[jacobolus]

In order for this device to qualify as ultimate it can't really use chords in order to stretch its usable number of inputs. That is a sacrifice in efficiency.

I don’t buy that. Chords are a great aid to efficiency if implemented carefully, especially if some chords are allowed to stand for common words/syllables/suffixes: I think a well-designed one-handed chording scheme with 10–15 switches could probably get up to 60–80 words/minute. With two-handed input and 20–30 switches total, I think we could get up to at least as fast as a standard keyboard (like 100–120 words/minute), or perhaps faster. The tricky part is figuring out how to optimize for movements that have fast transitions from one to the next, and use those for all the most common letters / blocks of letters.

[Input Nirvana]

I'm thinking for "ULTIMATE", a glove-shape frame, not an actual glove. Semi-ridgid so it has some "body" to use as a basis of positioning, yet be so comfortable you barely feel it. More of a partial "insert hand-fingers, then angle in". Like inserting into an "L" type movement. In then over.

MAJOR ISSUES:
1-Side to side movement of the middle fingers SUCKS. It's why I didn't care for my Datahand. I'm thinking this is the one drawback that could kill the device. The fingers may need to have more of a motion sensor at the knuckle rather than an actuation at the fingertip. I'm not sure yet how to get over this hurdle.

2-Chording, though it makes great sense in one respect, also doesn't make sense in the respect typing becomes a different animal once chording is introduced for the alpha characters. First attempt in mapping this out (in my mind) is to avoid chording for the alpha characters.

[vvp]

Side to side movement of the middle fingers SUCKS.

+1
Anything which requires this is not usable for me.
The reason I never even considered Datahand.

[hoggy]

Chording for single characters might be a pain, but chording for whole words (or even phrases) is quite different.  I'd be happy to have a chord for 'the','and' etc and to be honest, I'd be happy with chords for really unpopular letters such as q and z.

[katushkin]

WE HAVE THE TECHNOLOGY

Pretty sure finger extensions and retractions are the only movement that is repeatedly comfortable.

And as for not feeling it is there is the perfect frame of mind to go with I feel. When you sit or lie down, you are most comfortable when you don't even realise you are comfortable. Yeah, when you get into bed after a long day you let out a sort of sigh, but when you just drift off is the best.