Touch-down activation responds faster but cannot be cancelled

What it is

Touch-down activation executes when contact lands, responding faster but leaving no chance to withdraw by sliding out before lift. It suits low-risk, continuous, or minimum-latency feedback, not high-cost mistaken submission.

Why it happens

The system changes state or begins action directly on touch-down; later movement changes continuous parameters, not whether it occurred. What this saves is not an abstract "faster" — it is an entire lift-and-wait interval. Switching to touch-up means the action only happens the instant the finger truly leaves the screen, and that wait alone can eat a substantial share of the perceptible input-latency budget. For contexts where sound or effect must fire the moment contact happens — a musical instrument app's keys, a drum machine's pads — that wait is enough to make the response feel a beat behind, robbing it of the immediacy a percussive hit needs; people notice this clearly as "it's not keeping up with my finger," not just a vague sense of sluggishness.

Where it stops holding

Instant is not always better. When people expect an ordinary button to cancel on exit, touch-down breaks the learned model. Reversible previews, drag initiation, or volume adjustment can benefit, provided they are not mistaken for final commit. Instruments and drum machines, where the exact onset moment is highly perceptible, are among the few clear exceptions that must use touch-down, because response latency itself is a perceptible defect there and touch-up cannot meet that requirement. Conversely, everyday actions such as ordinary navigation or confirmation, which do not need that percussive immediacy, can save the same latency with touch-down, but that saved time rarely buys back the lost cancellability — the trade is not an even one.

Applying it

• Use only for low-risk, quickly reversible actions with a clear stop or undo.
• Show immediate continuous feedback that action has taken effect.
• How to check: measure the actual end-to-end latency difference between a touch-down and a touch-up implementation of the same gesture, and compare that concrete number against the acceptable latency ceiling for the context; only choose touch-down, and accept its non-cancellable cost, once the difference is confirmed to genuinely affect the experience.

Related

• Same group: C2.07.1 Touch-up activation allows changing one’s mind mid-action · C2.07.3 Irreversible actions must use touch-up activation · C2.07.4 A combination of preview on touch-down and commit on touch-up
• Nearby: C2.08 Touch cancellation gestures · D1 Output and feedback channels
• Search terms: touch down · latency · undo