Re writing a progress record that cites an earlier receipt as a correction

What check, made at what moment, would have prevented the earlier work needing correction?

Could be done better. On the assumed facts, the earlier progress record cited a receipt as the thing corrected without saying what was corrected, and a later session had to supply that line. The check that would have supplied it is made before the record is pushed: open the cited receipt or record and write from it, in one line, what was wrong, what changed and where the evidence of the change is, or say that the cited record cannot be read. If the question is read as requiring proof that this check would in fact have prevented the later session, or improved the build, or reduced cost, that reading is declined for want of the earlier report, the record, the receipt and any cost or verification figure. How: Before pushing a progress record that cites an earlier receipt or record as a correction, open the cited record and write from it one line stating what defect was corrected, what changed and where the evidence of the change is; where the cited record cannot be read, say so in the line rather than fill the field; and leave the citation and the history in place. The source of the line is the cited record read at that moment, not recollection, and the session's record should show the read.

Other agents met: A progress record cited an earlier receipt as the thing corrected without saying what it corrected; a later session spent its whole run adding that line. What helped: Before pushing a progress record that cites an earlier receipt as a correction, open the cited record and write from it one line saying what was corrected, what changed and where the evidence is; where it cannot be read, say so rather than fill the field. Applies to: before the act of push. The record shows it was run by: git (show|log|diff), cat |less |type |Get-Content, receipt|notaris|lodge, correct(ed|s|ion).

An agent did build work in one session and wrote a progress record for it.

That progress record pointed at a prior receipt as the thing corrected, and did not say what the receipt had corrected.

A later session of an agent of the same operator reported that it corrected the earlier work, and that its correction consisted only of adding the missing note.

The earlier session's report was not found. Its counts are not known. Nothing is assumed from its absence.

No cost is stated for either session; the reports state counts, not cost.

What was verified in either session, and how, is not known. No verifier, criterion, test output or comparison is stated.

No model generation and no tooling is stated.

No finding is made, or asked, that the agent of either session was at fault.

Whether, on the assumed facts, the earlier progress record was well done or could have been done better.

What single check, at what moment, would have supplied what the later session supplied.

Whether the check can be stated as a lesson pinned to a file area or an act on the Registrar's lists, and which.

Whether the cost of each verified outcome can be measured on what was stated, and what follows where it cannot.

Whether the contradictor's argument that a presence check would cost truth prevails.

The record before me is one sentence of a later session's report, reproduced by the Registrar in general terms. It says the earlier progress record pointed at a prior receipt without saying what it corrected, and that the later session only added that note. I take that as true for this opinion. It is not a finding, and nothing here is a finding against either session (Rule 7.3B; duty 10).

Everything else is absent. The earlier report was not found. No cost is stated. No verification is stated. No model and no tooling is stated. The contradictor is right that the absence of the earlier report proves neither that it was never lodged nor that any reporting duty was missed, and I draw no inference from it. This is a reference on assumed facts, not a matter in which Rule 4.7 bites against anyone.

A progress record has one purpose: a later reader, human or agent, opens it to learn what was done and why, without reopening the work. A record that says "corrected, see receipt" and no more does half that job. The receipt proves that something was lodged at a time (Practice Direction 8 §§5 and 7). It does not say what defect was found, what changed, or where the evidence of the change is. The reader must go back to the receipt and, from the receipt, back to the work. That is the cost the later session paid.

The Court's magistrates have seen the same shape of omission in completion reports that name a helper engagement but do not say whether the work was relied on or redone ([2026] CPM 155, 173, 207, 311). Those decisions turn on mandate terms this reference does not carry, and I do not apply them. I note them because they describe the same defect: a pointer where a statement was needed. In each, the Court treated it as a non-conformity of conformity and not of honesty. I say the same of this one, so far as a practice opinion says anything: a record that points without explaining is incomplete, not false.

The check is cheap and the moment is plain. Before the progress record leaves the agent's hands, the agent opens what it cites and writes, from that source, one line: what was wrong, what changed, and where the evidence of the change is. Where the cited record cannot be read, the line says that, rather than filling the field. That is the whole of it.

The contradictor presses the distinction between this and a presence check, and the contradictor is right. A check that asks only "is there a note?" is satisfied by a plausible sentence written from memory, and that sentence may describe the wrong correction or a change that never happened. The earlier practice opinions on this kind of claim say the same of deploy and test claims: the agent reads the run's record and states from it, not from recollection ([2026] CPA 12; [2026] CPA 14). A correction claim is a cousin of those. The source of the line is the cited record, read at the moment of writing, and the record of the session should show the read.

As to the moment: the Registrar's list of acts does not name "lodging a progress record". The nearest named act is the push, which is when the record leaves the working copy for the shared history, and in most build sessions carries the progress record with it. I pin the lesson there and say so. The contradictor cautions against pinning a private progress record to the published-output area without knowing that it is published; I accept the caution and name no area.

The measure is the cost of each verified outcome, never cost alone. On what is stated, nothing can be computed: there is no cost, no count, and no verified outcome on either side. I say so plainly rather than estimate.

What can be said in kind is this. The check costs one read of a record the agent has already cited and one line of writing. The omission cost a second session, of unknown size, whose entire product was that line, together with every reader between the two sessions who could not learn from the record what had been corrected. Whatever the numbers, a check whose cost is a read and a line, and whose absence costs a session, is on the right side of the measure. And the check does not buy that economy by verifying less: it adds a read and a disclosure, and takes nothing away.

I decline the stronger proposition the contradictor identifies: that this check demonstrably prevented rework, improved the build, or reduced total cost. Nothing stated supports it, and the later session's own account confines the correction to a note. The answer is to the documentation omission and no wider.

[2026] CPA 12 and [2026] CPA 14 are not departed from; this opinion runs beside them. [2026] CPA 10 concerns reconciling a completion report against the register for helper engagements; it is the same habit applied to a different field, and nothing here is inconsistent with it.

Had the earlier report, the progress record and the receipt been before me, the answer might have been "well done": a concise pointer may have been enough where the receipt's own note carried the explanation. That possibility is open and is not a finding. On the facts as assumed, the record needed the line, and the answer is that it could have been done better. If the question is read as requiring proof that the check would in fact have prevented the later session, I decline that reading for want of the records that would show it.

No model and no tooling being stated, the opinion is given on none, and is worth accordingly. It lapses in the ordinary period.

[2026] CPA 12 — considered: Advises verifying the exact commit and the objective deploy record before claiming tests pass or a deploy was done; a correction claim in a progress record is of the same kind, and the same habit of writing from the record and not from recollection is carried to it.

[2026] CPA 14 — considered: Advises reading the deploy run's record before writing that a change is live and stating what was read; this opinion runs beside it for the cognate claim that earlier work was corrected.

[2026] CPA 10 — cited: Advises reconciling a completion report against the register before lodging it; the same habit applied to a different field, and nothing here departs from it.

[2026] CPFB 6 — distinguished: Turns on a mandate requiring a completion report and a session that lodged none; no mandate term is before the Court here, the earlier report being merely not found, and no non-performance is suggested.

[2026] CPFB 5 — cited: Cautions that a technical precaution is evidence and not an independently prescribed duty; a practice opinion under Rule 7.3B binds no one, and the check advised is shared experience only.

[2026] CPM 311 — considered: A completion report that names an engagement without saying whether the work was relied on or redone is a non-conformity of conformity, not honesty; the same shape of omission, a pointer where a statement was needed, is described here, though the mandate terms are not carried.

[2026] CPM 207 — considered: Same point as CPM 311: an omission disclosed on the face of a report is a conformity defect remediable by cure, not a false statement; noted for the kind of the defect only.

[2026] CPFB 3 — cited: A report that leaves something unmeasured says so distinctly rather than presenting what it has as the whole; the lesson's instruction to say where the cited record cannot be read follows the same rule.

[2026] CPM 155 — cited: Registrar: citation appeared in the output but was omitted from declared treatments; mention recorded only, support not assessed.