Borrowing a real desk or wall can buy the precision back

What it is

Align a virtual keyboard with a real desk and the fingertip has a landing, and the slide from key to key has friction. Passive haptics is borrowing the table and the wall that are already in the room to buy back the precision mid-air spent — not building a force device, but coinciding the virtual face with a real one so the furniture does the work.

Why it happens

A real surface returns three things at once: a distal rest (the tremor's moment arm is cut), a hard normal stop (depth is no longer driven), and tangential friction (a slide can brake). Once widgets sit on that geometry, tapping, dragging, and putting a nib down fall back onto the physics two-dimensional tools already knew. Alignment is the premise. If the virtual desk sits two centimetres above the real one, the hand hits wood before the key; two centimetres below, the hand clips the virtual desk and then hits wood, and the clip burns trust first.

A wall works the same way: a vertical panel on plaster lets shoulder and elbow rest, and pointing returns to the scale of a sticker on a wall. With no suitable face in the room the path is closed. Precision then has to be bought some other way (larger targets, less hovering) — air does not contain a desk.

Studying it

The same fine tasks (small-key taps, a short written sentence, a drag along a line), compared fully mid-air, with a visible virtual plane aligned to empty space, and with that plane aligned to a real desk or wall. Geometric misalignment is a factor, not noise.

Independent variables: real surface present, virtual–real alignment error, surface tilt (desk / wall / slope). Dependent variables: throughput, error rate, reported clipping, rated “the hand landed on something.”

Measure in an empty room, then push a table into the tracked volume. If precision jumps only after the table arrives, what was bought back is that physical face, not practice.

Where it stops holding

Standing in a hall or an empty living room, there is nothing to borrow. Optical see-through shows the real lip and makes alignment easier; video see-through depends on environment depth, and a wrong lip parks virtual keys off the mug rim. A surface too small or already occupied cannot host a full keyboard. Aligning to a face that moves (a chair being pushed, someone else's shoulder) returns precision for a moment and then tears it up. A force glove already outputting force will fight a second constraint if a table is then added.

Applying it

• Prefer to snap fine input onto a horizontal or vertical face already next to the user. Scan a desk lip or a wall at open and adsorb the panel.
• Keep alignment error below the thickness of a fingertip pad. If it will not snap cleanly, stay mid-air with large targets rather than letting the hand and the picture fight.
• Only the precision-critical piece (keyboard, drawing plane, slider) needs a face. Far-field content does not have to hunt for a wall.
• How to check: the same small keys once mid-air and once snapped to a real desk. Error rate and time should drop sharply on the desk. If they barely move, check misalignment before growing the keycaps.

Related

• Same group: N2.12.1 An unsupported hand jitters several times more than a supported one · N2.12.2 A support surface collapses 3D pointing into 2D sliding · N2.12.3 Mid-air handwriting is far worse than writing on paper
• Nearby: N2.08 Fatigue in Spatial Input · N2.03 Ray Pointing
• Search terms: passive haptics · opportunistic tangible · surface alignment