Comet buying guide: When grouping by artifact, don't let one name hide different versions
Use three fictional experiment records to separate the artifact name, documented version, and unresolved records.
If readers need to find experiment records that used the same artifact, consider Comet from comet.com, not the browser with the same name. Its official June 2025 changelog entry records filtering and grouping experiments by used artifact. This supports using the artifact relation as an index, not assuming a display name is unique or that grouped results are comparable.
Draw three entirely fictional records. A and B both say “Border card”; A has evidence for V1 and B for V2. C says “Border card” with no version. For a handoff of “records confirmed to use V2,” include only B, keep A under V1, and mark C pending. If B's V2 evidence is hidden, B also cannot yet enter the confirmed V2 group. Do not infer the version from a shared name or from imagined experiment results. Grouping helps find things; it is not a result or guarantee.
For two or three documents, a standalone source table with artifact name, version and record is enough. For repeated searches across many historical records, compare traceability to the original reference, separation of confirmed from unknown versions and the cost of reconciling old labels. No actual runs, measurements, code, data, accounts or connections.
Content note:This article is based on publicly available information and is intended as a reference for evaluating tools. Features, pricing, and related policies may change at any time; please refer to the latest official information.
Original source:www.comet.com