Worked specimen / 02
Fictional teaching example
One mismatch. Eight fields.
The view is sorted by due date. The exported rows keep their insertion order.
Maple Board, its version, the observations, and all attempt counts below are invented for this exercise. This is not a report about a real product or a claim of a completed test.
Three rows are enough.
The input order differs from due-date order. That makes the two possible export behaviors easy to distinguish without a large dataset.
title,due_date,status
Cedar,2026-11-09,open
Alder,2026-11-07,open
Birch,2026-11-08,openDescribe the difference precisely.
{
"summary": "CSV export preserves insertion order after the board is sorted by due date",
"context": {
"product": "Maple Board — fictional training application",
"version_or_commit": "0.8.2 — fictional version",
"environment": "Fresh local training board; three invented cards; no customer records"
},
"preconditions": [
"Create Cedar (2026-11-09), Alder (2026-11-07), then Birch (2026-11-08).",
"Use the supplied three-row fixture; leave all statuses open."
],
"steps": [
"Open the board and select Due date, ascending.",
"Observe the visible order: Alder, Birch, Cedar.",
"Choose Export CSV without changing the filter.",
"Open the downloaded file and read its row order."
],
"expected": "If export is specified to preserve the active sort: Alder, Birch, Cedar.",
"observed": "In this fictional specimen, CSV rows are Cedar, Alder, Birch, while the board remains Alder, Birch, Cedar.",
"reproduction": {
"attempts": 3,
"failures": 3,
"control": "In this invented exercise, an unsorted board and its export share insertion order.",
"scope_unknown": "Export sort requirements, other versions, other environments, and other sort keys have not been established."
},
"evidence": [
{
"file": "fixture.csv",
"purpose": "Invented input only; included with this reference."
},
{
"kind": "observation",
"note": "All outcomes and attempt counts on this page are fictional teaching data. No real product test or attached failure capture is claimed."
}
]
}A repeatable difference needs a contract.
The fictional result supports a narrow statement: the view and the export have different orders under the listed conditions. It does not establish that the export is incorrect.
First check whether exports are supposed to preserve the visible sort. If the contract says insertion order, this may be a clarity problem in the interface. If it says active sort, the same report becomes a focused defect report. If there is no stated contract, record the ambiguity rather than silently choosing one.
A useful next check changes one condition: choose another sort key, or compare one earlier version. Record the result even if it contradicts the first explanation. Do not turn one small fixture into a claim about every export.