A file list can look orderly while answering the wrong question. Upload time tells a user something about an event in the system. It does not by itself establish which drawing was approved, whether a revision has been superseded or whether the document is suitable for a particular purpose. I learned that distinction through an account I approved at Ternavell Infrastructure, where a convenient latest-file convention became a stronger documentary claim than its evidence supported.
In the fictional 2009 case, the project-information account contained ninety-six floor plans selected using upload order. A review against the revision register identified twenty-eight as superseded versions. Evan Tarswell challenged the assumption, and I supported A$480000 of review and correction. We retained approved revision references, source dates and document status. The exercise did not certify the building, establish as-built conditions or prove that every unflagged drawing was current for every possible use.
Separate a file event from a documentary conclusion
The original convention was attractive because it was easy to explain. A user could open the newest item and assume it was the relevant one. Maintaining that apparent simplicity required little discussion of the distinction between a file arriving and a document acquiring approval. I had accepted the convenience without adequately examining the claim it made. That was my responsibility, even though the files themselves had not been altered by the selection rule.
An older version can be uploaded again. A newer file can exist without the approval status a reader expects. These possibilities do not mean that upload timestamps are useless; they mean the timestamps answer a narrower question. I want the account to retain the event information while locating approval in the source that actually establishes it. The mistake comes when the system’s easy ordering becomes a substitute for the document owner’s evidence.
Evan returned to the revision register rather than debate which file looked most current. That shifted the discussion from presentation to authority. We needed to know which revision the responsible process recognised, and for what purpose, instead of treating recent arrival as a universal signal of suitability. The data team could maintain that relationship and expose uncertainty. It could not create an approval merely by assigning a current label to the chosen file.
I would also distinguish the document’s status from the condition of the physical place it describes. An approved drawing is part of a documentary account. Whether the building matches it, whether a later change occurred and what inspection may be required are separate questions for the relevant owners. A public case about document control should not borrow the authority of engineering or safety work that it did not undertake. The clarity of a version list does not close those questions.
Correct the selection without overstating the review
The twenty-eight superseded plans were a specific finding in the reviewed set of ninety-six. It would have been easy to turn the remaining sixty-eight into a statement that all other drawings were verified current. That conclusion would exceed what the account establishes. I wanted the report to say which records the exercise flagged and how their status was supported, without converting the absence of a flag into unlimited assurance about every possible use.
The correction retained the approved revision reference, source, date and status. We did not simply sort the same files in another way and ask users to trust the replacement list. The useful change was the relationship between the selected document and the evidence supporting its role. Original records remained traceable so another reviewer could understand what had changed and challenge the revised account if the source evidence required a different interpretation.
There is a reasonable objection that requiring users to inspect approval detail for every document could make ordinary work cumbersome. I agree that the interface should support an efficient selection. The answer is to make that selection depend on a defensible status account and provide the supporting detail where needed. A simple view can rest on careful evidence. It becomes unreliable when the ease of the view is achieved by replacing the evidence with an assumption that happens to be convenient.
The A$480000 expenditure supported review and correction. I would not turn that amount into a quantified avoided-loss claim or describe the programme as proof that every future version would be correct. The delivered output was a more defensible document account and a clearer explanation of the status used in selection. The work mattered because it withdrew an unsupported implication that could influence a user’s reliance on the file, not because the case supplies a dramatic physical outcome.
Make the status understandable after the file leaves its folder
Document information is often reused outside the interface in which it was selected. A plan may be downloaded, forwarded or referred to in another report. The relationship to the approved revision can become less visible at that point. I want the institution to consider how users will identify the source and status in the context of the intended workflow, rather than assume the original folder order will continue to explain what the document represents.
This does not require a data leader to prescribe every engineering-document process. The appropriate arrangements belong to the institution and responsible professional owners. My concern is that the public account of the work should identify the evidence maintained and its limitations. The data team supports the relationship between records; it should not imply that it independently approved the design or inspected the asset merely because information is presented coherently.
For a board or senior executive, the useful question is what a statement such as current drawing actually means in the report being reviewed. Who determines that status, which evidence supports it and what has the review not examined? The board does not need to manage each file. It should be able to understand the scope of the account before treating a tidy list as an assurance about the wider project or property.
All organisations, people and documents described here are fictional. A real narrative would need verified evidence and permission to discuss potentially restricted material. My argument is also an account of my own error: I allowed a file event to stand for a documentary conclusion. Correcting it required more than a new list. It required restoring the link to approval evidence and leaving the physical and professional questions visible for the owners whose work could actually answer them.