Rendered from docs/decisions/0078-a-recorded-terminal-demonstration-may-show-a-frozen-number-behind-a-recorded-on-date-marker.md in the Headwater
corpus. Every document on this half of the site is typed by the taxonomy
the descriptor names: corpus.json.
A recorded terminal demonstration may show a frozen number behind a recorded-on-date marker
Context
#601 asks what holds a recorded terminal demonstration true against a real run. HW-DR-0061 answered that question by confining a recording to the tutorial's scratch corpus. Nothing in this repository refreshes a number frozen inside a binary asset. That record left one question with the product owner: whether principle 11 of spec 0 may relax for a marketing asset. HW-DR-0061 named the question and left the rule in force.
A later build-order run put the question to the owner directly, as RULING #601. May a recording show output counts a live run does not reproduce? The condition offered was a visible marker of the recording's own age. The owner answered on the issue's own timeline: "Relax the rule with a recorded-on-<date> marker." A recording may show output counts a live run does not reproduce, if it carries a visible recorded-on-<date> marker (comment).
This is a ruling on the one asset HW-DR-0061 confined. It answers the owner's question with disclosure, and not with the enforcement HW-DR-0061 named as the reopening condition. This record is the successor HW-DR-0061 needs: no decision record in this corpus had yet revised a current decision in place.
Decision
A front-page recording may show a number a live run does not produce, if a visible marker stands beside it. The marker takes the exact form recorded on YYYY-MM-DD, as plain text next to the embed. A reader who sees the recording also sees the date it stopped matching a live run. The marker is what tells them so.
The marker discloses the freeze. It does not hold the frozen number against anything. Nothing regenerates the recording. Nothing compares its numbers against a later run. HW-DR-0061 named that comparison, with a recorder on the runner, as the condition to reopen this question. The owner chose the marker instead. The comparison HW-DR-0061 described is an accepted deviation for this one asset now, not a gap this corpus still owes.
HW-DR-0061's still-standing findings carry forward unchanged. The tool is vhs. The embed format is an animated GIF. GitHub renders no asciinema player in a README, and it blocks an animated third-party SVG in its image proxy. This record narrows the rule that confined a recording's subject matter. It leaves the choice of tool and format exactly as HW-DR-0061 stated them.
What a blocking check may hold is the marker's presence, and not the frozen number's accuracy. A GIF is a binary no fixture reads. The check that enforces this ruling reads the README's text instead. It fails when a recording embed carries no recorded on YYYY-MM-DD marker beside it, and it never opens the GIF.
Consequences
The recording clause of HW-OBL-0183 is narrowed by this record. That record's discharge condition, a CI step that regenerates a committed binary asset and compares it against the tree, was written against HW-DR-0061's confinement rule. For a front-page recording, that condition does not describe what closes the #601-shaped half of the obligation. The owner chose a marker over a compare-and-regenerate step, and the marker's presence is what the blocking check holds now.
A recording of a corpus number, once committed, ages by construction and stays true by disclosure rather than by refresh. A reader who runs the command named in the recording, and sees different numbers, has not been shown a lie. The marker beside the recording already told them the count was recorded on a stated date, and not read from their own run.
Nothing here mints a recording. No recorder is installed on the host that measured this record. The committed recording and its embed stay open on #601 until a recorder exists. This record answers only what a committed recording may show, and how it discloses its own age, once one is built.