Robotics reporting often compresses three different moments into one visible timestamp: the date a source was published, the date something actually happened in the field, and the date a record was later revised. On a news-oriented site that covers robotics events and technical material, mixing those moments can quietly distort a story even when every individual fact is correct. A correction workflow that separates them protects readers and keeps the archive usable.
A single "published on" line is convenient, but it is not the same as the event itself. The W3C PROV Model Primer explains that provenance can describe entities, the activities that generate or change them, and the agents responsible for those activities, so a document is not only a static object but also the result of processes that happened at particular times (PROV Model Primer). In robotics reporting, the entity may be an article, the activity may be a field demonstration, and a later activity may be an update to that article. Treating the publication date as the event date collapses the activity that generated the original observation with the activity that generated the page.
DCAT Version 3 similarly distinguishes between a dataset as an abstract resource and its distributions as accessible forms, which helps show why a catalog entry is not merely a file with a date (Data Catalog Vocabulary (DCAT) - Version 3). Recommended documentation practice is to record, for each robotics item, the observed event window, the source document date, and the later revision date as separate fields rather than one merged date.
A vendor note, conference page or institutional release may appear weeks after a demonstration, prototype run or public showing. The PROV primer notes that different perspectives capture agent-centered, object-centered or process-centered provenance, such as who took a picture or which steps produced a chart (PROV Model Primer). In a robotics context, that means the person operating a system, the team that wrote the release and the newsroom that republished it are separate agents with separate responsibilities.
When a reporter writes about a manipulator arm performing a pick-and-place routine, the routine happened on its own day. The press material describing it may be published later, and the site article later still. Recommended documentation practice is to keep the event date in the lede or a dated caption, while the source publication appears in a provenance note linked to the underlying material. Readers should never have to infer the event date from the page date alone.
Revisions create new entities. The PROV primer gives the example that a second version of a document generated by translation is a new entity produced by an activity (PROV Model Primer). A robotics article corrected after a vendor clarifies a specification is likewise not the identical record it was before the edit.
DCAT 3 says the revision supports describing versioning of resources and distinguishes the abstract dataset from its manifestations (Data Catalog Vocabulary (DCAT) - Version 3). For editorial work, that supports keeping a visible changelog: what changed, when, and which earlier claim was refined. Recommended documentation practice is to label an edited piece with both the initial publication date and the last material revision date, so the timeline of the record is transparent without confusing the original event.
Consider a hypothetical example. Suppose a robotics newsroom published a short item on 12 March 2026 about a warehouse mobile robot demonstration, using a source document dated 5 March 2026 that described an event said to have occurred on 28 February 2026. The page initially showed only "12 March 2026" at the top. Later, on 20 April 2026, the vendor corrected the demonstration date to 3 March 2026 in a follow-up note, and the newsroom updated the article.
The problem in this hypothetical example is not false reporting of the robot's function, but date conflation. Readers searching for what happened in late February could miss or misplace the item. The fix is hypothetical but practical: keep 28 February 2026 in the first version only if sourced, then correct to 3 March 2026 with a note that the event date was revised on 20 April 2026, while the source publication of 5 March 2026 and the page publication of 12 March 2026 stay in a provenance block.
A usable checklist begins with three distinct timestamps. First, the observed event date or window, taken from the source description of what happened. Second, the source publication date, which is when the referenced document was made available. Third, the page publication and revision dates for our own article. Recommended documentation practice is to place these in a small metadata line or footnote, not to merge them into a headline date.
Next, identify the agents and activities. Who demonstrated or operated the system, who authored the source, and who edited our page? The PROV primer's view of agents and responsibility supports recording that an organization may act through people and software, and that provenance can express chains of responsibility (PROV Model Primer). For a robotics newsroom, the editor correcting the date is an agent of a later activity, separated from the field team that produced the first observation.
Finally, mark distributions and versions. DCAT 3 treats a distribution as an accessible form of a dataset and allows catalogs to include services and series, so different copies of the same material can be described (Data Catalog Vocabulary (DCAT) - Version 3). Recommended documentation practice is to note whether the cited source is a PDF, an HTML page or an archived copy, because the same correction may appear in several distributions with different access paths.
On a site that maintains both industry updates and technical guides, the workflow scales by item type. A short industry notice may need only event date, source date and update note. A longer technical piece may need the demo day, the datasheet publication, the interview used for context and the later correction, each shown so a reader can see which fact came from which moment.
The discipline is small but durable. Do not let the page date stand in for the event date. Record the source separately from the occurrence, record the revision separately from both, and keep a changelog when a date is corrected. That practice keeps robotics reporting accurate over time without claiming that any cited standard requires our exact checklist; it is our recommended documentation practice built on the provenance and catalog concepts in the two W3C sources above.