A stylus moves smoothly while rendering occurs once per animation frame. The parent pointer event contains four coalesced samples with timestamps up to the dispatch time and three predicted samples beyond it. When the real path bends, later confirmed samples disagree with the prediction. A robust renderer keeps the confirmed history, replaces speculative pixels and never rewrites a prediction as a measured point.
This lesson continues The Gold Standard of Websites, the Gold Standard of Education and the How X Works hub. Every stylus trace, coordinate table and latency calculation is an original teaching fixture. The lesson does not claim that every device supplies the same sample rate, that predictions are measurements, or that one browser’s smoothing and horizon should be copied as a universal algorithm.
Record parent event type, pointerId and pointerType, trusted status, timestamp basis, client coordinates, pressure, tilt and buttons, coalesced and predicted list lengths and ordering, display frame time, device and browser versions, transforms, correction method and which pixels remain speculative.
The W3C Pointer Events Level 3 Recommendation defines coalesced and predicted event lists, their ordering, coordinates and relation to pointermove and pointerrawupdate. The standard intentionally leaves prediction and coalescing algorithms to user agents, so the worked traces separate normative event meaning from measured implementation behaviour.
THE GOLD STANDARD · WEBSITES
Coalesced points describe measured history; predicted points estimate a future path
A dispatched pointermove or pointerrawupdate can represent more device samples than the browser chose to dispatch individually. getCoalescedEvents() exposes confirmed samples associated with that event. getPredictedEvents() exposes user-agent estimates for positions expected after the parent event. They can support the same ink stroke, but they have different evidential status and must be stored, drawn and corrected differently.
Fast route
Start with the decisive rule.
Start this route
Measurement route
Build the browser state and event ledger.
Start this route
Comparison route
Separate the nearest lookalikes.
Start this route
Transfer route
Change the context and retest.
Start this route
Reporting route
State limits and continue reading.
Start this route
Chapter contents
Open the full chapter index
- 1. Give the history-versus-forecast rule
- 2. Understand why coalescing happens
- 3. Trace the parent event
- 4. Preserve time order
- 5. Keep coordinates in one space
- 6. Use coalesced points for confirmed ink
- 7. Render predictions as temporary
- 8. Measure prediction error
- 9. Handle empty lists
- 10. Separate pointermove and pointerrawupdate
- 11. Respect pressure and geometry
- 12. Avoid duplicate parent points
- 13. Control event-listener work
- 14. Test changed motion
- 15. Repair common shortcuts
- 16. Answer practical questions
- 17. Build a reproducible trace
- 18. Use the Recommendation and continue
1. Give the history-versus-forecast rule
Treat coalesced events as confirmed samples the user agent grouped with a dispatched event, and predicted events as estimates after that event. Never place both into one unlabeled point array.
Original worked example. Parent event at 120.0 ms yields coalesced samples at 116.0, 118.0 and 120.0 ms, plus predictions at 122.0 and 124.0 ms. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. The stylus turns at 121 ms; the next confirmed sample diverges from both predictions. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Store source=coalesced or predicted with every point and preserve timestamps. If this check fails, mark the claim unavailable and repair the method before continuing.
Keep platform state, application state and the final claim in separate fields so a convenient interface label cannot rewrite raw browser evidence.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
2. Understand why coalescing happens
Input hardware can sample faster than event dispatch or rendering. Coalescing reduces dispatch overhead while retaining access to intermediate motion.
Original worked example. A device supplies eight samples during one 16.7 ms frame; the browser dispatches two parent events whose coalesced lists cover the observed path. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. Lower the display rate while holding input rate steady and compare list lengths. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Measure rather than infer device rate from one event count. If this check fails, mark the claim unavailable and repair the method before continuing.
Use a positive case, a clean negative case and an unavailable case. Unavailable evidence is not a denial; it means the field could not be established.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
Previous chapter · Contents · Next chapter
3. Trace the parent event
The parent event is the anchor for its lists. Its pointerId, type and dispatch context determine where the samples belong.
Original worked example. Two pens move at once. Each parent pointermove returns only the list for its own pointer identity. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. Interleave touch and pen input and verify grouping remains per pointer. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Key every trace by pointerId and sequence number, not by arrival time alone. If this check fails, mark the claim unavailable and repair the method before continuing.
Change one consequential premise while keeping the rest fixed. Predict the new branch before inspecting it and preserve any disagreement.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
Previous chapter · Contents · Next chapter
4. Preserve time order
Coalesced samples are ordered chronologically, and predicted timestamps extend beyond the parent event. Sorting all points without source labels can create a false measured future.
Original worked example. Three confirmed timestamps precede two predicted timestamps. A naive global sort looks smooth but hides the boundary. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. Add a late processing delay; timestamps remain the evidence of sample time rather than callback time. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Record both event.timeStamp and processing time with a documented clock basis. If this check fails, mark the claim unavailable and repair the method before continuing.
Carry one plausible alternative through the state and event ledger, then choose the next observation that separates it from the leading account.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
Previous chapter · Contents · Next chapter
5. Keep coordinates in one space
Client, page, screen and transformed canvas coordinates are different. Predictions and confirmed samples must pass through the same declared transform before comparison.
Original worked example. A scaled canvas doubles device-independent coordinates; only confirmed points are transformed in the buggy renderer, creating an apparent prediction error. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. Apply the same DOMMatrix to both lists and recompute residuals. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Archive raw client coordinates, transform matrix and rendered coordinates. If this check fails, mark the claim unavailable and repair the method before continuing.
Name the document, element, event, pointer, item or permission object involved. Similar visible results can come from different platform paths.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
Previous chapter · Contents · Next chapter
6. Use coalesced points for confirmed ink
A drawing app can connect coalesced points to preserve detail that one parent coordinate would omit. The result remains bounded by supplied samples and coordinate precision.
Original worked example. The parent endpoints form a straight chord while four coalesced samples reveal a small curve. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. Disable coalesced access and compare path length and curvature error. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Keep the raw sample sequence and do not oversell smoothed interpolation as observed motion. If this check fails, mark the claim unavailable and repair the method before continuing.
Archive source, fixture, options, events, timestamps, browser version and raw exceptions so another reader can reproduce the conclusion.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
Previous chapter · Contents · Next chapter
7. Render predictions as temporary
Predicted points can reduce perceived latency when drawn speculatively, but they need a visible or internal status that allows later replacement.
Original worked example. The app draws a faint two-point tail beyond the last confirmed sample, then replaces it on the next event. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. Force a sharp corner; stale speculative pixels are cleared before confirmed ink is committed. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Track a reversible predicted layer or segment range rather than painting permanently. If this check fails, mark the claim unavailable and repair the method before continuing.
Keep platform state, application state and the final claim in separate fields so a convenient interface label cannot rewrite raw browser evidence.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
Previous chapter · Contents · Next chapter
8. Measure prediction error
Accuracy is established only when later confirmed samples can be matched to earlier predictions on time and coordinate definitions.
Original worked example. Prediction at 124 ms is compared with the confirmed sample nearest 124 ms, producing a 6-pixel residual. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. Change the matching tolerance and report how the residual summary changes. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Publish matching rule, time tolerance, coordinate space and unavailable matches. If this check fails, mark the claim unavailable and repair the method before continuing.
Use a positive case, a clean negative case and an unavailable case. Unavailable evidence is not a denial; it means the field could not be established.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
Previous chapter · Contents · Next chapter
9. Handle empty lists
A browser or device may return no coalesced or predicted events. Empty evidence is a supported outcome, not permission to fabricate samples.
Original worked example. Mouse movement yields one parent event and empty prediction list; the renderer uses the parent coordinate. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. Switch to a high-rate pen and observe nonempty coalesced data. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Feature-detect methods, record list lengths and keep a parent-event fallback. If this check fails, mark the claim unavailable and repair the method before continuing.
Change one consequential premise while keeping the rest fixed. Predict the new branch before inspecting it and preserve any disagreement.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
Previous chapter · Contents · Next chapter
10. Separate pointermove and pointerrawupdate
The event types have different dispatch and performance characteristics. A trace must name which parent produced the lists.
Original worked example. pointerrawupdate arrives more often than pointermove in the tested environment, but both can expose coalesced samples. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. Remove the raw listener and compare workload and path fidelity. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Record listener type, options, event count, processing cost and frame misses. If this check fails, mark the claim unavailable and repair the method before continuing.
Carry one plausible alternative through the state and event ledger, then choose the next observation that separates it from the leading account.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
Previous chapter · Contents · Next chapter
11. Respect pressure and geometry
Coordinates are only part of a pointer sample. Pressure, tilt, width and height may vary and can require their own interpolation or fallback rules.
Original worked example. Two coalesced points share position but change pressure, affecting brush width. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. Use a mouse whose pressure convention differs and keep pointerType visible. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Store all consumed fields and document default or unsupported values. If this check fails, mark the claim unavailable and repair the method before continuing.
Name the document, element, event, pointer, item or permission object involved. Similar visible results can come from different platform paths.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
Previous chapter · Contents · Next chapter
12. Avoid duplicate parent points
Implementations can make the parent event correspond to the final coalesced point. Blindly appending both may duplicate a coordinate and distort velocity.
Original worked example. The last coalesced timestamp and coordinate equal the parent event; the naive path contains the point twice. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. Deduplicate by explicit identity rule, then test a case where values match coincidentally. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Document whether the renderer consumes the list, parent or both and how equality is judged. If this check fails, mark the claim unavailable and repair the method before continuing.
Archive source, fixture, options, events, timestamps, browser version and raw exceptions so another reader can reproduce the conclusion.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
Previous chapter · Contents · Next chapter
13. Control event-listener work
Heavy synchronous processing can raise latency and alter observed dispatch cadence. The measurement system must not become the main cause of missed frames.
Original worked example. Logging every property blocks 12 ms per event and grows list lengths. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. Buffer compact records and process them after the frame; compare counts and latency. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Measure handler duration, dropped frames and buffer loss alongside input samples. If this check fails, mark the claim unavailable and repair the method before continuing.
Keep platform state, application state and the final claim in separate fields so a convenient interface label cannot rewrite raw browser evidence.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
Previous chapter · Contents · Next chapter
14. Test changed motion
Prediction quality depends on motion continuity. Straight travel, stopping, reversal and corners are distinct transfer cases.
Original worked example. Predictions are close during steady motion and overshoot after an abrupt stop. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. Repeat at slower speed with the same geometry and compare residuals. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Use scripted or repeatable paths where possible and retain the uncertainty of human trials. If this check fails, mark the claim unavailable and repair the method before continuing.
Use a positive case, a clean negative case and an unavailable case. Unavailable evidence is not a denial; it means the field could not be established.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
Previous chapter · Contents · Next chapter
15. Repair common shortcuts
Coalesced does not mean synthetic, predicted does not mean confirmed, more points do not guarantee lower latency, and a smooth line does not prove accuracy.
Original worked example. A blinded table mixes source labels and asks which points may be committed permanently. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. Hide coordinates and use timestamps alone to identify the forecast boundary. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Require source, time, pointer identity and correction state in every conclusion. If this check fails, mark the claim unavailable and repair the method before continuing.
Change one consequential premise while keeping the rest fixed. Predict the new branch before inspecting it and preserve any disagreement.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
Previous chapter · Contents · Next chapter
16. Answer practical questions
Use confirmed points for the durable stroke, predictions for reversible latency masking, and a parent-event fallback for empty or unsupported lists.
Original worked example. An accessible drawing tool stores confirmed vectors, shows a temporary prediction tail and offers an undoable stroke history. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. Disable prediction and verify drawing remains correct though less anticipatory. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Audit touch-action, keyboard alternatives, pointer capture and performance budget. If this check fails, mark the claim unavailable and repair the method before continuing.
Carry one plausible alternative through the state and event ledger, then choose the next observation that separates it from the leading account.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
Previous chapter · Contents · Next chapter
17. Build a reproducible trace
A useful trace includes raw lists, frame times, transforms, rendered segments and the rule that reconciles predictions with later samples.
Original worked example. Ten events produce a table of parent, three confirmed and two predicted points each, plus residuals on the next frame. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. Change the display refresh rate and rerun without rewriting the interpretation rule. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Archive code, device, browser, screen rate, timestamps and output images. If this check fails, mark the claim unavailable and repair the method before continuing.
Name the document, element, event, pointer, item or permission object involved. Similar visible results can come from different platform paths.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
Previous chapter · Contents · Next chapter
18. Use the Recommendation and continue
Pointer Events Level 3 gives the lists their distinct meaning while leaving algorithms and availability to user agents.
Original worked example. The final report cites the Recommendation and dates measured coalescing and prediction behaviour. The objects, values and observations are explicit so the reader can reconstruct the decision instead of trusting a label.
Changed-context transfer. When hardware, browser or transform changes, rerun the saved straight, corner and stop fixtures. State what changed, predict the consequence and revise only the conclusion that depended on that condition.
Practical evidence check. Save exact source, access date, support matrix, raw event JSON, matching rule and correction evidence. If this check fails, mark the claim unavailable and repair the method before continuing.
Archive source, fixture, options, events, timestamps, browser version and raw exceptions so another reader can reproduce the conclusion.
Before closing, reconcile the original fixture against the chapter rule: identify the relevant objects, platform state, event order and missing evidence; change one condition and mark the first unsupported inference. Preserve the raw trace beside the revised result. End the chapter with three sentences: what the evidence establishes, what it does not establish and what observation would change the conclusion. This keeps the lesson useful when the browser, document state or user context changes. A strong record also names the datum, denominator, section or coordinate system that governed the decision, so a later reader can repeat the reasoning without importing an unstated shortcut.
Grouped reading routes
Core route
Companion lessons
- Separating navigator.canShare() Validation from navigator.share() Outcomes
- Separating Fullscreen Request Promise Resolution from fullscreenchange Event Delivery
- Separating Clipboard read() Items from readText() Plain Text
Related specialist lessons
Sources and further study
W3C Pointer Events Level 3 Recommendation. Used for the standard definitions, material mechanisms or method limits developed in this lesson.
The worked examples and transfer checks are developed here and state their own assumptions. The linked sources support the underlying web-platform mechanisms, state and method limits.
Return to contents · Return to The Gold Standard of Websites