22. White Space sensitivity and measurement reference

Use Settings → White Space & solver sensitivity when the default profile produces evidence that does not match your publication's measure or house style. Change one relevant setting, apply it, inspect a representative scope again, and compare like-for-like results. A lower count after raising a threshold is a policy change, not proof that source text improved.

Setting	Units / permitted range	Meaning and practical effect
Minimum river span	3–12 lines	Required corridor length. Raising it suppresses shorter alignments; lowering it admits shorter patterns for review.
Maximum horizontal drift	0.1–2 em	Allowed movement between neighbouring corridor spaces relative to local type size. Raising it admits more sloped/variable paths.
Short final line below	5–40%	Maximum ending width relative to the widest non-final line in the same column. Narrow columns use a conservative ceiling of 10%.
Minimum paragraph lines	3–12	Minimum complete paragraph length for visual short-ending review; it is not a widow/orphan definition.
Rag variation threshold	15–70%	Normalized variation of eligible non-final left-aligned line widths. Raising it requires stronger variation before review.
Terminal column below	10–75%	Terminal line-count ratio for comparable captured containers. It does not determine whether an intentionally short final column is wrong.
Allow subtle glyph scaling in solver	On / Off	Permits eligible solver alternatives involving horizontal glyph scaling. This checkbox does not itself alter InDesign.

The basic stored White Space policy uses 15% short-ending width, three minimum paragraph lines, 35% rag variation and a 50% terminal-column ratio; an active profile or saved customization can differ. The built-in profile summary is the practical starting point. Advanced internal solver limits and score weights can also be carried in validated profiles; do not look for a visible control that the current panel does not expose.

Composition details and the score

Open White Space → Composition and choose a module's Details to read its metrics. Review and the count chips return to the one shared findings queue. Opening or closing Details does not rescan, mark or create a second result set.

The combined composition score requires White Space and Styles in the same accepted run. It is a 0–100 composition indicator, not the percentage of all document errors found. Focused scans, missing evidence and partial nested-text coverage can prevent a valid combined score. Do not compare the score across changed profiles, scopes or module selections without stating that change.

Module states are also specific: Not scanned means there is no accepted snapshot; Not included means White Space was excluded; Update required means the run is stale; Unavailable means the required retained module evidence is missing. Clean means there are no open findings in that fresh module summary, after the applicable local review decisions. It does not assert that no findings were originally detected, that dismissed items were repaired, or that omitted structures are error-free.

Rivers Details shows detected paths and High + Critical / Medium / Low counts. Under Professional details the selected finding can expose confidence, continuity, impact, angle and residual. These measurements describe detector evidence; they are not a probability that an editor must change the text.

Texture measurements

Choose a page in Texture preview to inspect its density cells. The preview stays inside the panel. Large grids can be display-sampled; this is a rendering simplification, not a claim that the source was rescanned at lower accuracy.

Measurement	Correct interpretation
Estimated density	Area estimated from word bounding boxes, not exact rendered glyph outlines.
Rhythm	Leading/baseline consistency in comparable story, style and size groups; Not assessed is not zero consistency.
Column alerts	Captured physical columns classified as overdense.
Density variance	Variation in estimated texture; a smaller value alone does not certify a better design.
Whitespace entropy	Distribution complexity of inter-word whitespace, not a standalone quality grade.
Rag / balance assessed and skipped	Which eligible comparisons ran and which were excluded.
Empty physical columns	Observed empty containers when topology is captured; partial/unavailable topology is stated explicitly. Editorial intent is not inferred.

Copyfit measurements

Overset stories counts affected threaded stories; Overset characters counts their hidden text. Widows identify final paragraph lines isolated at a later frame/column top; Orphans identify first lines isolated at an earlier frame/column bottom. Short endings are a separate visual measure and do not require a boundary crossing.

Under Professional details, read tight/loose lines, audited stories, short-ending coverage and skipped reasons. Sibling-column depth compares occupied depths of sibling columns in the same frame; it is not a left-versus-right facing-page balance score. A partial paragraph, protected role, unsupported alignment or unavailable reference measure can be excluded rather than assigned a false clean result.