Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mehen.ophi.dev/llms.txt

Use this file to discover all available pages before exploring further.

Tables are valuable up to a point. A table with 6–60 cells usually improves comprehension; a table with 300+ cells is usually a maintenance artifact that belongs in generated output or structured data.

​
Per-table burden

Combines wide, long, and cell-count saturation terms with missing-header, empty-cell, and alignment-complexity penalties: 
TableBurdenScore = 0.5 · mean(T_burden) + 0.5 · max(T_burden)
The blend of mean and max means a single very large table moves the score even when most other tables are fine.

​
Table Scaffolding Score

Uses a piecewise size credit:
CellsCredit
1–50.2 — too small to matter.
6–601.0 — useful comparison scaffold.
61–150max(0, 1 − (cells − 60) / 120).
> 150More burden than scaffold; credit is near zero.
A hard warning fires when cells > 300 OR cols > 12 OR rows > 100, with the suggested remediation: split the table, generate it from structured data, or move the source to YAML/JSON/CSV.

​
How DMI uses it

Table Burden contributes to DMI via the T_norm term; Table Scaffolding is one of the components of Good Scaffold.

​
References

  • Tufte, E. R. (2001). The Visual Display of Quantitative Information, 2nd ed. Graphics Press — classic reference on when tables aid comprehension and when they become noise.
  • Few, S. (2012). Show Me the Numbers: Designing Tables and Graphs to Enlighten, 2nd ed. Analytics Press — concrete guidance on table sizing and the cell-count thresholds adopted here.
  • W3C: Tables Tutorial — Tables in WCAG — defines header-row requirements that the missing-header penalty enforces.

​
See also