For search-and-highlight revision in a long manuscript, use pass one to discover candidates and pass two to build the context needed for a later decision. Pass one records locations and chapter counts without editing the manuscript. Pass two captures the sentence before, the candidate sentence, the sentence after, the speaker, and the viewpoint. The result is a complete evidence packet for another editor.
This article stops before classifying candidates as preferred forms, intentional exceptions, or pending decisions. It does not apply approved replacements or perform the final reconciliation. Its role is candidate discovery and context preparation.
Editing during search changes the candidate population
Suppose you search for the repeated expression “suddenly” and delete the first result immediately. The count decreases and later result positions move. If you add “something” to the search list halfway through, the final record no longer represents one stable discovery scope.
Judgment can also drift. An editor may remove aggressively in chapter one and retain similar cases late in the session. The resulting chapter distribution then mixes characteristics of the manuscript with changes in the editor’s threshold.
Do not alter the text during discovery. Pass one ends when every occurrence in the declared scope has an ID—not when its literary value has been decided.
Pass one collects locations and chapter counts
Freeze the terms and scope. In this example, search for “suddenly” and “something” in copies of three chapters.
Assign a candidate ID to every occurrence, such as D-001 and D-002. Record:
- candidate ID;
- search term;
- filename and chapter ID;
- scene heading;
- a short quotation of the target sentence; and
- occurrence number within the chapter.
Do not mark an occurrence for deletion, retention, or author review. If you use highlight colors, let blue mean the “suddenly” search group and orange mean the “something” group. The colors in the external candidate sheet identify search sets; they do not represent correct and incorrect prose.
At the end of the pass, total each chapter. If chapter one has eight occurrences of “suddenly,” chapter two has two, and chapter three has seven, preserve that distribution without declaring chapter one bad. It can guide the order of context collection in pass two.
Pass two builds a context card for every ID
Do not add or remove search terms in pass two. Open each ID from the pass-one list and collect the information a later editor will need.
Place the preceding sentence, target sentence, and following sentence on one context card. Add the speaker for dialogue and the viewpoint character for narration. When necessary, record the referent or time within the scene as an observed fact.
For example, D-014 may contain “He suddenly looked toward the window.” Capture that an alarm sounded in the preceding sentence, that he notices light outside in the following sentence, and that he is the viewpoint character. Do not yet decide whether “suddenly” should stay. Make the trigger and consequence available to the later decision-maker.
An isolated target sentence can make a transition cue look like a verbal habit or make deliberate ambiguity look accidental. The context pass exists to prevent that evidence loss.
Turn chapter distribution into a candidate map
Once the cards exist, map counts and locations by chapter.
Look beyond totals. Record whether occurrences cluster at chapter openings, appear mainly in dialogue, follow one viewpoint character, or concentrate within a few nearby paragraphs.
Eight occurrences scattered across eight scenes differ from seven occurrences concentrated in one conversation. Preserving that difference reduces the risk that a later decision treats every match identically.
Use observational language only: “eight in chapter one,” “four in SC12,” or “six in the protagonist’s viewpoint.” Do not label a cluster excessive, wrong, or exceptional during discovery.
Check completeness with IDs and counts
Compare the pass-one list with the pass-two cards. If the IDs run from D-001 through D-017, there should be seventeen context cards, and the chapter totals should also sum to seventeen.
If a card is missing, return to context collection rather than editing the manuscript. If pass two exposes a candidate that pass one did not capture, record it as an additional discovery together with the search condition and reason. Do not silently append it as though the original scope had included it.
The completion package contains term-specific candidate lists, a chapter-distribution map, and a context card for every ID. Post-replacement counts and exception totals are not completion criteria for this stage.
Use Rune Studio search as the discovery entry point
Rune Studio’s feature materials document search within a file, across a workspace, and within selected files, folders, or paths. This is documentation evidence, not evidence of the current screen or a completed search operation. Those scopes are relevant when considering single-file discovery, whole-project discovery, and discovery limited to selected project items.
Documentation that selected text can become a search term is not the same as documentation that the selected passage can become the search target. This article claims neither documentation nor screen verification for the latter. It limits the candidate target scope to the documented file, workspace, or selected file, folder, or path level.
That scope does not establish automatic judgment that “suddenly” is a bad habit or that “something” is unacceptable ambiguity. Rune Studio does not become the owner of preferred, intentional, or pending classification in this article. Candidate IDs, colors, distribution maps, and context cards are editorial records.
Replacement is not the center of this workflow. Stop after collecting stable search results and context. Keep the documented search scope separate from the illustrative count of seventeen, which is not a measured product result.
Hand decisions and replacement to STUDIO-339
STUDIO-339 takes prepared spelling candidates and decides preferred form, intentional exception, or pending status in a decision queue. It also owns the later application and reconciliation of resolved rows.
This article hands over candidate ID, term, location, chapter distribution, surrounding context, speaker, and viewpoint. It does not attach a “should change” verdict. Keeping discovery separate prevents the candidate rule from shifting while decisions are being made.
Complete a two-term candidate packet
Begin with only “suddenly” and “something.” In pass one, assign an ID to every occurrence across three chapters and total the chapters. In pass two, add preceding, target, and following sentences plus speaker and viewpoint to every ID.
When the number of IDs, the sum of chapter counts, and the number of context cards match, the discovery packet is complete. The manuscript remains unchanged. A packet ready for the separate decision stage is the result of this two-pass search-and-highlight workflow.
If you need file-, workspace-, or selected-item search scopes for discovery, check the current scope on the Rune Studio product page and begin with two terms in copies of three chapters.
