Move a fisheye's focus and every element re-arranges, so the target you reached for slides away

What it is

The fisheye's warp field is bound to the focus: move the focus and every element re-arranges — the target that sat under the pointer slides away, and by the time the user reaches for it, it is somewhere else. This "chase and miss" effect is the fisheye's most notorious usability trap, converting a statically usable view into a moving fairground.

Why it happens

The drift comes from the self-referential structure of "pointer position defines the warp field": in hover-driven fisheyes the pointer is the focus, yet the pointer's position is itself determined by where elements currently sit — the user moves toward the target, the target relocates in the re-warping field, and on arrival it has moved again. In control terms this is an unstable loop, worst near the boundary between compression and focus zones. Three mitigation families exist: decoupling (the focus lags the pointer, or the focus moves by click/keyboard instead of hover), locking (press on the target to freeze the warp before confirming), and attenuation (a gentler magnification gradient shrinks the displacement). All three serve one principle: the aiming act itself must not move the target.

Where it stops holding

Drift severity tracks the interaction binding: hover binding is the worst, explicit focus-setting milder, keyboard or scrollbar-driven mildest. Touchscreens lack hover and so dodge the worst form, but the warp at the instant of tapping still shifts the target under the finger — touch fisheyes must sample the touch point before the warp applies. Drift also scales with warp strength: the steeper the magnification gradient, the larger the displacement per unit of pointer travel; turning the strength down is the cheapest mitigation there is.

Applying it

• Hover fisheyes need focus lag or damping (the focus smoothly trails the pointer rather than binding instantly).
• Provide a two-stage selection — "click to lock the target → warp freezes → confirm click" — instead of direct clicking.
• Verification: time users clicking specified small targets inside the fisheye and count attempts; a failure rate clearly above the plain view means drift is uncontrolled — add locking or reduce the gradient.

Related

• Same group: U6.07.1 Fisheye achieves focus plus context by local magnification and peripheral compression · U6.07.2 Geometric distortion breaks positional encoding; magnitudes stop being readable · U6.07.4 Distortion views suit navigation and location, not reading or comparison · U6.07.5 A way to switch the distortion off and return to the plain view must exist
• Nearby: U6.06.5 The focus position must be movable by the user · U6.09.2 Touchscreens have no hover and need alternative paths
• Search terms: fisheye pointing · target drifting · interaction stability