Separate event, publication, and research dates in filings
A filing timeline needs separate fields for the event, public disclosure, and your inspection. Never use a download timestamp as evidence of when an event occurred or when other investors could have known about it. The three clocks answer different questions and can differ even when they appear on the same calendar day.
The 8-K cover identifies the earliest reported event date, while HKEXnews results display release-time information. These examples show why a source can contain more than one relevant date. The three-clock record is a suggested research convention, not a claim that every venue publishes identical timestamp fields.
Preserve the original before normalizing
For each date, write what it describes and where it appears. A signature date describes signing. A document-cover date may identify the report rather than publication. A platform release timestamp describes that platform’s record. Your access time describes when you inspected it. Do not pick one field and rename it “date” across the whole archive.
Preserve the source’s time zone when stated. If only a date is available, keep date-level precision; do not invent midnight to make a spreadsheet sort. A normalized timestamp can be useful for ordering across markets, but store it beside the original with the conversion assumption visible. If the zone is unknown, leave the exact order unresolved.
Hypothetical example across two markets
Assume fictional Larch Ports signs an agreement on October 5, releases a notice on October 6 local time, and you inspect it on October 8. The scenario and dates are hypothetical.
| Clock | Hypothetical field | Meaning |
|---|---|---|
| Event | October 5; date only | Signing described in the notice |
| Publication | October 6; platform local time | When this source shows release |
| Research | October 8; your recorded zone | When the document was checked |
| Coverage cutoff | October 7 for an earlier memo | Earlier memo may include publication but not your later inspection |
A retrospective memo should not say “verified on October 7” when your first inspection happened on October 8. It can say the source was publicly released earlier, if the source record supports that, while clearly identifying the later verification date.
Build the timeline around questions
For “what did management do?” use the event clock and preserve uncertainty about effectiveness or completion. For “what was publicly available?” use the disclosure record and the defined cutoff. For “what did our research check?” use your inspection log. Keeping these questions separate prevents accidental claims about historical knowledge.
When two reports discuss the same event, retain both publication records. Similar text released in different markets may have different timestamps or additional attachments. Do not overwrite the first publication field with whichever document you downloaded most recently. Connect the documents to one event while preserving their separate release histories.
Verification and stopping rules
Reopen the notice and identify the sentence establishing the event date. Then inspect the platform record for publication metadata. Finally, compare these with your own access log. If a field came from a secondary source, identify that provenance rather than presenting it as platform metadata.
The timeline is sufficient when another reader can tell which date supports which claim and where ordering remains uncertain. It cannot prove the first moment any person learned information privately. Nor does it determine legal filing timeliness without the relevant rule, exceptions, and calendar. Avoid drawing those conclusions from a simple timestamp comparison.
Common errors include converting an unknown time zone, treating a dated agreement as an effective agreement, and attaching later corrections to an earlier information set. A precise “date only; ordering unresolved” note is more accurate than a fabricated time that looks computationally convenient.
FAQ
Should every date be converted to UTC?
Only when the source precision and time zone support conversion. Keep the original value and label assumptions so readers can audit any normalized field.
Can a later download prove earlier public availability?
The download alone cannot. Use the publication record for availability and your download or inspection record for your own research activity.