Hover previews where the stroke will land before contact
Aliases: stylus hover · hover preview · tracking state
What it is
Hover (proximity tracking) is the state in which the nib has not touched the screen but has entered the digitizer’s near field. The system can draw a cursor, brush footprint, or snap hint so people see where the stroke will land. It is the tracking state in three-state input, between out-of-range and contact dragging. Hover previews the upcoming contact location; it is not the parallax caused by glass thickness after contact.
Why it happens
An EMR or active electrostatic pen still couples several millimetres above the glass. Field strength falls nonlinearly with distance; firmware reports hover coordinates—often with a coarse height—between a contact threshold and a leave threshold. The interface draws that coordinate as a non-permanent preview (cross, dot, brush outline) and only then deposits ink on contact. Aiming becomes a visual loop instead of a proprioceptive guess at the nib. Without hover, the first sample is known only at contact, so fine targets are luck or undo. The preview must share the contact point’s calibration; if the hover cursor and the ink disagree, people learn to ignore the preview and the channel is wasted. When height is small but contact has not happened, the preview should already be stable, or the last millimetres of jitter will be painted into the first stroke.
Studying it
Compare visible hover, no cursor, and a cursor deliberately offset from the contact point. Tasks: small-target taps, landing in a gap between ruled lines, and calligraphic starts on a grid.
Independent variables: hover visibility, preview glyph (dot / footprint), systematic preview-to-contact offset. Dependent measures: contact endpoint error, pre-contact corrections, whether the start hits the gap, and “I can see where it will land”.
Log the difference between hover coordinates and the first contact sample; that is whether the preview is trustworthy.
Where it stops holding
A passive capacitive stylus, a finger, and protocols with near field off have no reliable hover; aiming cannot be built on a preview. If hover height is too short, the preview arrives too late to close the loop. Making hover a hot zone that fires before contact (menus opening in tracking) turns the state into false activation. Thick glass or thick film leaves the near field intact but changes how the preview relates to the visual nib; that belongs with parallax calibration, not with merely switching a cursor on.
Applying it
- Before contact, show a cursor or brush footprint in the same coordinate frame as the ink; on contact, that footprint becomes ink, not a second location.
- Keep the preview non-destructive: no document change, no focus grab, no panel covering the landing point.
- Default hover preview on for precise starts (marquee, text insertion, vector anchors).
- How to check: aim at a point smaller than the visual nib. People should park the preview on the point before contact, and the first ink should coincide with it. Turn the preview off as a control; error should rise.