Effective Content Units (ECU) - mehen

ECU normalizes review mass so that a small-but-dense ADR is not overshadowed by a long linear README in volume comparisons. It blends prose word count with weighted contributions from code, tables, diagrams, math, and raw HTML/MDX.

Formula

ECU = W / 240
    + 0.35 · CLOC
    + 0.06 · table_cells
    + 0.40 · diagram_nodes
    + 0.25 · diagram_edges
    + 0.12 · math_tokens
    + 0.20 · raw_html_or_mdx_lines

The W / 240 term anchors on the standard adult silent-reading-rate scale (~240 words per minute). Other coefficients are derived to make a small, dense diagram contribute roughly as much as the equivalent paragraph of explanation.

Interpretation bands

ECU	Meaning
`< 5`	Small.
`5–20`	Normal.
`20–60`	Large.
`> 60`	Documentation subsystem — likely wants a split.

Why not just word count?

Word count systematically undercounts artifact-heavy docs (architecture diagrams, big tables, code-only references) and over-weights prose-heavy docs that may be relatively easy to skim. ECU was introduced so “dense vs. linear” shows up directly in volume comparisons.

References

Trauzettel-Klosinski, S. & Dietz, K. (2012). Standardized assessment of reading performance: the new International Reading Speed Texts IReST. Investigative Ophthalmology & Visual Science 53(9): 5452–5461 — the silent-reading-rate evidence base that anchors the W / 240 term. DOI.
Brysbaert, M. (2019). How many words do we read per minute? A review and meta-analysis of reading rate. Journal of Memory and Language 109: 104047. DOI.

See also

LOC family — components that feed ECU.
Section tree — per-section ECU aggregation.