Check whether issue reports describe the same failure

Two issue titles can describe the same visible symptom while requiring different investigations. Compare the attempted task, starting conditions, and observed sequence before calling reports duplicates. Use the expressly fictional document example below to build an issue-equivalence record. Its purpose is to preserve relevant differences, not to diagnose an actual Kiguri defect or claim a customer problem was solved.

Normalize the reports without replacing their originals

Keep each original reference and write a compact comparison: intended task, location or version, starting state, steps, actual result, and available evidence. Mark missing details as unknown. Do not make the reports look equivalent by dropping conditions that appear inconvenient.

In fictional report R1, a reader follows a saved old-title link and sees Unavailable. The document appears in the ordinary unfiltered list, and its new-title link opens. In R2, a reader opens the correct folder with the Assigned to me filter active and sees an empty list. The same new-title link opens directly. Both titles say document missing, but their routes and visible failures differ.

Match the conditions that matter

Use synthetic fixtures with the same intended document and fictional workspace version. Record the active filter and exact route, rather than saying both readers opened the application. Repeat each sequence as described before changing it. A comparison made after silently clearing the filter cannot explain the second report's original result.

For R1, compare the old and new links while keeping the list state unchanged. For R2, compare the filtered and unfiltered lists while keeping the direct link unchanged. In the fixture, replacing the old bookmark resolves the unavailable route, while clearing the assignment filter reveals the document in R2. These are separate controlled observations; they do not establish a shared underlying defect merely because both readers eventually find the document.

Compare triggering behavior and resolution scope

Ask whether one canonical reproduction represents both reports. A sequence beginning with the stale bookmark reproduces R1, but does not exercise R2's filtered-list condition. Likewise, a filter change cannot repair an outdated bookmark. Keep separate investigations unless further evidence establishes a common failure that explains both sequences.

Add a third fictional report, R3: another reader follows that same old-title link, gets Unavailable, and can open the new-title link from the unfiltered list. R3 is a stronger duplicate candidate for R1. Compare any remaining differences, such as browser context, before linking them. A matching symptom is a lead; equivalent supported conditions and a common investigation target justify the proposed link.

Preserve distinct evidence when linking is justified

Write the equivalence decision and its reason in the record. For R1 and R3, retain both report references, their observed contexts, and the old-link reproduction. Keep R2 separately, or label a relationship as unverified if investigation remains open. Linking does not mean deleting inconvenient evidence from a canonical issue.

If a proposed correction later addresses the old route, check that the canonical description still represents each linked report. Do not mark the filtered-list report resolved solely because another reporter confirms their bookmark now works. Each accepted reproduction needs its own relevant completion evidence.

Reopen equivalence when new evidence conflicts

Ask another reviewer to map each report to the canonical steps and expected observation. If they must invent a filter state or omit a differing result, revisit the duplicate decision. Record what remains unknown and who will investigate it.

The finish check is a traceable answer to why these reports share one failure investigation, or why they remain separate. A smaller issue list is not automatically a better one. Preserving the conditions that matter avoids making a clean-looking list at the expense of a problem nobody has actually examined.