Manuscript versioning: reconcile the text, citations and revision letter

1. Separate file revision from publication stage

NISO's published RP-8-2008 vocabulary distinguishes submitted, accepted and published article stages, among others. A team's file label such as S2 is a separate revision identity. Approval by the coordinating author does not make S2 an accepted manuscript.

The NISO working-group page read on 15 September 2026 describes the later draft as under review. Cite the dated published recommendation for the terms used here, rather than calling a draft a final replacement. This guide supplies an original workflow, not a claim that a particular editor or software service follows it.

Reference: NISO RP-8-2008: journal article versions · NISO: Journal Article Versions revision working group

2. Keep the request and submitted package together

MV42 begins with submitted manuscript S1 and reviewer request R1. Preserve the exact S1 text, its reference file and the request before starting revisions. A copy of a draft made later is not necessarily the file the reviewer read.

Create a change map keyed to the reviewer comments. For R1.2, record the requested clarification, the base location, the proposed edit and the person who must assess its meaning. A suggested change can be declined with reasons; the map must not assume that every reviewer request will be adopted.

3. Trace the branches before combining them

Record each branch's base. If a contributor edited a different starting copy, compare that copy first; otherwise an apparent deletion may simply be text the contributor never received. Keep the returned branches even after reconciliation.

On small screens, scroll the table horizontally, or focus it and use the arrow keys.

4. Reconcile meaning and references across documents

The coordinator takes the methods clarification to its scientific owner, then checks the Arabic wording, table footnote and response letter against the same decision. A language improvement can be retained while its outdated definition is corrected. Agreement on fluent wording alone is insufficient.

For C7, compare the claim with the intended source and record which source version supports it. Do not replace the reference simply because another entry is newer or has a similar title. Until that check is complete, the citation and the letter's claim remain open.

On small screens, scroll the table horizontally, or focus it and use the arrow keys.

5. Verify the exact package rather than the filename

Once the C7 check is resolved in an actual project, record its outcome with the manuscript changes. MV42 stops before that verification: its package remains not ready for resubmission. A coordinator cannot close the package by renaming a file or ignoring the unresolved reference.

The release list should identify the manuscript, Arabic text where separate, response letter and reference file that belong together. A checksum can distinguish saved bytes, but does not establish their scientific meaning. Review the rendered tables, references and response locations as well as the source documents.

6. Close the handoff with recipient readback

After the remaining checks are completed in a real workflow, ask the receiving coordinator to identify the exact package and open its key documents. Record whether the supplied files, reference version and response locations match the agreed list. A sent link is not evidence that those checks happened.

Preserve the submitted package and the corresponding receipt separately from the next working copy. If a later error is discovered, retain the earlier evidence and follow the journal's correction process. This exercise performs no upload, software recovery or real recipient check.

Questions for a practical final check

• Is the reviewed S1 package preserved with R1?
• Does each branch identify its base?
• Do EN, AR, tables and the response letter express the same decision?
• Is C7 still explicitly unresolved until checked?
• Is coordinator approval separate from journal acceptance and receipt?