Core terms
Content Comparison
The full reviewed source-to-screen comparison. This is the default public view and the source for every optional reader lens.
Canonical Dataset
The one shared body of source facts, comparison records, evidence, mappings, and review state for an Adaptation. Public views pull from this dataset rather than storing separate versions of the same research.
Comparison Record
One structured comparison about a specific subject such as a character, relationship, event, theme, setting, structural choice, or ending element. It keeps the source detail and screen detail together.
Reader Lens
An optional presentation of the canonical comparison records: I Watched the Drama, I Know the Source, or I’m New to Both. A lens changes emphasis and spoiler handling; it does not create separate content.
Change types
Kept
The element remains substantially recognizable in both versions.
Changed
The same element exists in both, but its execution, detail, motivation, function, or outcome differs.
Added
The screen version contains an element with no meaningful source counterpart.
Removed
The source contains an element that does not meaningfully appear on screen.
Merged
Multiple source elements become one screen element.
Split
One source element becomes multiple screen elements.
Expanded
An element exists in both, but the screen version gives it more space, prominence, or function.
Condensed
An element exists in both, but the screen version compresses it.
Reordered
An element remains but appears in a different sequence or point in the story.
Reworked
The underlying idea remains recognizable while the mechanism, function, relationship, or meaning changes substantially.
Likeness
Likeness describes how recognizable an element remains across versions: Near Same, Close, Recognizable, Reworked, Minimal, None, or Unknown. It is descriptive, not a fidelity score.
What each version gives you
A comparison record may identify a factual element or experience available in one version that is absent or meaningfully different in the other. This is not a recommendation or a judgment of quality.
Spoiler scope and level
Records identify whether a spoiler belongs to the source, screen version, both, or neither, and how significant it is. When spoiler protection hides part of a set, public counts can show visible / total so the reader can see that more reviewed material exists without revealing it.
Evidence
A research reference used to support or verify a source fact or comparison claim. Evidence can be reviewed independently from the comparison record it supports.
Review state
Proposed has not been decided. Accepted is approved for the canonical comparison. Needs Work requires more research or editing. Rejected is intentionally excluded from the accepted comparison while remaining part of the project audit trail.
Confidence
Verified means the comparison is supported and reviewed. Probable preserves a supported but not fully confirmed conclusion. Unresolved means the claim still requires better evidence or review.
Completion
A project-level measure of whether the current canonical research has been reprocessed, reviewed, verified, and cleared of blocking issues. It is not a fidelity score and is not increased simply by adding more records.
Reader Correction
A public correction report that can be routed into the project's review workflow so a claim can be researched and resolved without creating a second body of comparison data.