STUDY 001 / PROPOSED METHOD

Useful progress.
At what cost?

Can progress updates adapt to their own cost without making the wait feel longer?

A / DIAGNOSTIC PILOT

Measure the cost first

Compare the renderer’s normal intermediate updates with a final-frame-only diagnostic. Record snapshot work, visible feedback, and final completion. The diagnostic is a measurement control, not a proposed user experience.

B / CONDITIONAL STUDY

Adapt the update cadence

If the pilot exposes a reproducible cost, compare an adaptive update policy with fixed intervals. Keep the requested page and final image quality identical, while tracking the gaps between useful updates.

The evidence

Measure the first useful frame, final render completion, visible feedback gaps, and snapshot overhead. A completed render task is not a paint timestamp: inspect visual updates and browser traces separately. Check final output correctness and interaction responsiveness.

The controls

Pin PDF.js 6.5.438 and verify whether each run actually uses its renderer worker. Record fallback runs separately. Use identical documents, display settings, final quality, and interaction tasks; repeat serial trials in balanced order. Keep file acquisition outside the rendering comparison.

A useful failure

Stop after the pilot if updates add no reproducible cost. An adaptive policy also fails if it delays useful feedback, creates longer unexplained pauses, or produces no improvement over a simple fixed interval. No benchmark runs have been collected yet.

Prior work

Progressive drawing and worker rendering already exist. This proposal examines the cost of the update policy, without claiming to invent progressive PDF rendering.

Measurement and publication standards