HCI Compendium — Theories, Models, and Metrics of UI/UX

draft

The Koder Stack's conceptual foundation for interface quality: cognitive laws, heuristics, score metrics/models, visual design & computational aesthetics, predictive interaction, and accessibility. Grounds the UI Quality Score (stack-RFC-030).

Why hci and not ui/ux? The content is the field of Human-Computer Interaction — spanning UI, UX, usability, ergonomics, perception, cognition, interaction, and metrics. "UI" would be too narrow; "UX" is the synonym-of-practice (a subset). The slug is precise (RFC-024 separates slug↔title); the title carries the friendly UX/UI label. Decided via /k-arch (2026-06-29).

This compendium is the conceptual source of the Stack's interface quality index (the UI Quality Score, stack-RFC-030). Each chapter favors what is measurable (formula, threshold, validated score model), because the ultimate goal is a computable index, not a collection of opinions. Grounded research (Claude Code, 6-agent fan-out, 2026-06-29); sources cited in each chapter.

Chapters

# Chapter What it covers
1 UX laws & cognitive principles Fitts MT=a+b·log₂(2A/W), Hick RT=a+b·log₂(n+1), Miller 7±2, Tesler, Doherty 400ms, Weber–Fechner, Von Restorff, peak–end, aesthetic–usability… — 17 laws with formula, measurability tag, and proxy.
2 Heuristics & evaluation methods Nielsen 10, Shneiderman 8, Norman, Gerhardt-Powals; severity 0–4; heuristic evaluation / cognitive walkthrough → the JUDGE tier; 15-criteria checklist.
3 Quantitative metrics & scoring models SUS, UMUX-LITE→SUS (0.65·x+22.9), SUPR-Q, SEQ, HEARTGSM, ISO 9241-11, ISOIEC 25010, SUM, NPS — the math of each + how to compose an index.
4 Visual design & computational aesthetics Measurable Gestalt, typographic scale, 8pt grid, WCAG (L1+.05)/(L2+.05) + APCA, Ngo's 14 measures (0–1), Hasler–Süsstrunk, Rosenholtz, AIM.
5 Interaction & predictive HCI Norman (gulfs of executionevaluation), affordance×signifier, *OMSKLM*(predicts seconds), Fitts/Hick, information scent, error taxonomy.
6 Accessibility & standards WCAG 2.2 (POUR, AAAAAA) with 18 AUTO checks and thresholds, ARIA/accname, Microsoft Inclusive Design, COGA, ISO 9241-110.
7 Synthesis — the UQS algorithm Derives the 0–100 index: SUM (standardize→average) + SUPR-Q (sub-factors) + ISO 9241-11 (pillars) + AUTO/JUDGE tiers + profiles + gates + Sauro–Lewis bands.

How to use

  • Concept/theory → chapters 01–06 (citable reference).
  • Build/calibrate the index → chapter 07 + stack-RFC-030 (norm) + meta#387 (implementation).
  • Compendium maintenance/k-compendium hci (config in _search.md).

Related: stack-RFC-030 (UQS — norm) · specs/naming/brand-score.kmd (sibling index, for names) · policies/architecture-quality.kmd (sibling rubric, for architecture) · specs/themes/verge.kmd R8 (AA contrast — inherited gate).