Parallel operations cannot be estimated by adding time

What it is

Perception, memory retrieval, motor preparation, and system waiting in real interaction tasks often overlap or run in parallel. Adding every theoretical stage assumes strict seriality and overestimates total time. Completion time follows dependencies and a critical path, not the arithmetic sum of all activity.

Why it happens

People can keep looking at a target while moving a pointer, prepare a next step during animation, or begin another process while one runs automatically. Only stages with information or resource dependencies must wait. Overlap depends on task, device, feedback timing, and expertise, so timelines contain intersections rather than neat serial blocks.

Studying it

Reconstruct a task timeline with gaze, trajectory, key, UI events, and recording to identify actual onset and offset. Change the duration of one process and observe whether total time changes by the same amount; if not, processes overlap. Use critical-path or process modelling rather than adding segment means.

Where it stops holding

Parallelism does not mean unlimited resources: high cognitive load, occlusion, one-handed input, or safety confirmation can force serial processing. Assuming too much overlap also underestimates complex work. Determine overlap from actual dependency and behavioural data, not mere theoretical possibility.

Related

• Same group: B1.17.1 Laws calibrated under separate conditions double-count preparation and response when added · B1.17.3 Adding group-average predictions amplifies rather than cancels error · B1.17.4 A predicted difference smaller than individual variation is not design evidence · B1.17.5 A law's ordering of alternatives is more credible than its absolute value
• Nearby: B1.01 Fitts's law · B1.04 Hick-Hyman law
• Search terms: parallel processing · critical path · task-time model