The usable zone is the intersection of view and reach

What it is

A hip pouch is in reach but only enters the headset picture after a look down; a switch on the far wall is in view but beyond the arm. Direct manipulation lives in the visual-manual workspace: seen and reachable. Either volume alone is half a workspace.

Headset cameras add a third, forward-facing tracking cone. A hand behind the waist can still reach and still be unseen, and tracking usually dies with it. The usable set is the intersection of three volumes, not the sphere swept by arm length.

Why it happens

Unaided direct touch needs three facts at once: the target is in the current image, it sits inside the arm's kinematic envelope, and the hand is inside the tracker's cone. Consumer binocular overlap is roughly ninety to one hundred and ten degrees, far smaller than the solid angle an arm can sweep — hip, back, and behind the ear are reachable and off-screen. Conversely a button metres away can sit in the middle of the picture and still be an empty grab.

People compensate by turning the target into view or walking it into reach. Both cost, so frequent actions spontaneously collapse into a patch in front of the eyes. The tracking cone is narrower and more frontal than the eyes: a hand at the waist or beside the head may still be half-visible while the skeleton is already gone. The third layer is stricter than “I can see it.”

Studying it

Lay targets around the body in azimuth and distance, split into in-view and in-reach, out-of-view but in-reach, in-view but out of reach. Log discovery time, head turns, and whether the person switches to a ray. Tag tracking loss separately; it is not a miss of seeing.

Independent variables: target azimuth, distance, headset field of view, whether turning or walking is allowed. Dependent variables: discovery latency, head-motion amplitude, fraction of tracking loss, whether the finish used a direct grab or a far-field point.

Lock heading and run the set again: the “must turn to use” slice of the intersection appears on its own. A study that only plants widgets dead ahead never measures this.

Where it stops holding

Ray and gaze-plus-pinch do not need the target inside arm length; the intersection constraint almost drops for far-field selection. Seated in a chair that cannot yaw, the out-of-view reachable slice is nearly dead — heading travel is small and walking is zero. Video see-through makes it harder than optical see-through to catch the real hand at the hip in peripheral vision. A wide industrial headset grows the intersection, but if the tracking cameras still face forward, the third layer does not grow with it.

Applying it

• Put directly grabbed widgets in the intersection of the current view and the forward arm envelope: about chest height, a little below the line of sight, within an arm.
• Hip and back are for low-frequency objects found by proprioception. If the grab must be seen, add a “summon to the front” action.
• Do not design beyond-arm targets as direct grabs, or offer another selection method until the person has walked in.
• How to check: place a frequent control where the arm can reach it but the view cannot. Count discovery time and head turns. If every use starts with a look down or a side turn, it is not in the intersection — move it, rather than growing its hit target.

Related

• Same group: N2.10.2 Cost is asymmetric across the body midline · N2.10.3 Upper bound is set by the shoulder; lower bound by the cost of bending · N2.10.4 Reach travels with the person; a world-fixed panel can be walked out of
• Nearby: N2.06 Reach Space · N3.01 Locking Modes
• Search terms: visual-manual workspace · reachable workspace · interaction volume