An unresolved revision is a result worth preserving

A digital archive can look orderly while making a claim its source does not support. Every record has an identifier, every field contains a value and the current revision appears easy to find. That presentation is useful only if the status has a defensible basis. When historical material contains conflicting dates or unclear approval marks, filling the field may require a judgement that belongs to the record owner. A conversion team should not acquire that authority simply because the delivery schedule expects a complete index.

We encountered the issue in a 12,000-record sample of industrial drawings. Reviewers could not confidently determine the revision status of 720 items from the material supplied. A production lead proposed taking the latest visible date as the deciding rule. It was a practical way to finish the field and might have been appropriate if the authorised owner had specified it for that record set. In the circumstances, it was an unapproved inference. I asked for the uncertain items to remain explicit and for the client to provide the necessary classification basis.

## Separate readable evidence from authorised status

The task begins by identifying what the team can actually observe. An operator may read a date, a revision letter or a note. That observation does not automatically establish which document is approved for a particular engineering use. Different source conventions may exist within an archive, and a later date can relate to a copy, an annotation or an administrative action. The operating process should record the evidence available and the question it leaves open, rather than ask staff to make the archive internally consistent by treating unlike markings as if they meant the same thing.

This distinction is not an argument against using rules in metadata work. Consistent rules are essential for scale. The question is who authorises the rule and what it is intended to establish. If the client defines a classification approach, the conversion team should understand and apply it within the agreed scope, retaining the basis for exceptions. If the source falls outside that approach, the team needs a route back to the owner. A rule that works for most records should not become a reason to conceal the cases it does not answer.

Our reviewers therefore used a visible unresolved category for the 720 questions. The category needed more substance than a blank field. It identified the source reference and the matter requiring clarification, so that the client could act on it. We also explained what the status meant in the delivery account. The company was not declaring the drawing invalid or unusable in every context. It was saying that the conversion service could not support the proposed revision classification from the evidence and authority then available. That narrower statement was both more accurate and more useful.

## Make the clarification obligation part of the plan

Exceptions take time on both sides of a service relationship. A provider can prepare a clear question, but only the relevant client owner may be able to answer it. If that capacity has not been considered, the programme can become stuck while both organisations believe the other is delaying completion. I would now agree the expected clarification route and response planning before a similar conversion begins. Historical archives often contain ambiguity; treating every such question as an unforeseeable event makes the schedule less credible than acknowledging the dependency from the outset.

In our case, the client programme manager was concerned about presenting an unfinished archive at an internal review. Our commercial lead was concerned about extra scope. We agreed a bounded exercise with ₹11 lakh additional operating expense and nine working days. That did not purchase authority for us to decide the engineering status. It funded the preparation and handling of questions, together with the client's review process. The distinction kept the commercial discussion connected to the work being performed, rather than imply that a higher price could make the underlying uncertainty disappear by itself.

The client's designated owners resolved 540 of the 720 questions. The other 180 remained open with their source references. We did not silently apply a default to close them when the review period ended. The unresolved population was part of the result, and the client needed to understand its implications. A defined clarification exercise can make meaningful progress without answering every question. If a report demands universal closure as the only acceptable outcome, it creates pressure to replace a useful exception with an unsupported certainty, precisely when careful judgement is most needed.

## Report the remaining uncertainty without disguising progress

The arithmetic should be clear. In the original sample, 11,280 records did not require the clarification route. Of the 720 that did, 540 were resolved and 180 remained uncertain. Those categories describe the work without implying that all 12,000 drawings had been certified by Thervani. A resolved metadata status meant that the designated owner supplied a basis for that classification. It did not transform the conversion provider into an engineering approver. The report should allow a reader to understand both the progress and the limit without reconstructing the categories from several different percentages.

Usability depends on carrying that distinction into the delivered system. If the exception is visible only in a handover document while the search interface presents the item as current, the company has not preserved the practical meaning of its own review. The client and provider need an agreed way to display and handle unresolved status within their respective responsibilities. That can affect retrieval, user instructions and later updates. The precise solution depends on the service and system, but the principle is simple: a limitation that matters to use should remain available where the record is used.

Directors overseeing a records-dependent service should ask whether management knows which judgements it is authorised to make. The question extends beyond drawings. An organisation can infer a status in order to complete a process, then present that status as if it came from the proper authority. Boards should examine the escalation route, the treatment of unresolved cases and the commercial incentives surrounding closure. They should also resist the urge to decide the underlying technical question themselves. Oversight should establish that responsibility is properly assigned and evidence preserved, not move the same unsupported inference into a more senior meeting.

I regard the 180 unresolved records as a result worth retaining, not a blemish to remove from the presentation. They identify work that still belongs to the authorised owner and prevent a user from mistaking our completion pressure for engineering assurance. The next programme should plan the clarification capacity sooner and make the delivery treatment clearer from the start. But a better process will not always eliminate uncertainty in historical material. The responsible service is one that can finish its own work while leaving the source's remaining limits visible, specific and assigned to someone able to address them.