Description
Estimates interface consistency as the length of the shortest rule set or compressed description that generates its interactions; shorter descriptions mean a more learnable, consistent grammar.
Kolmogorov complexity is the length of the shortest program that produces a given output. It is uncomputable in general, but it can be approximated by compression or by the minimum description length principle. Applied to interfaces, consistency can be framed as compressibility: if a single rule (for example 'swipe left always deletes, with undo') covers many cases, the interaction grammar is short; if each screen has its own exceptions, the shortest description is long. The method encodes interactions as a formal grammar or action-response table, counts rules and exceptions, or compresses the encoded table, and compares versions. Lower description length predicts faster learning and fewer errors, echoing task-action grammar from early HCI. It provides a quantitative consistency metric for design systems and audits.
Attributes
- Discipline
- Information Theory, Physics & Mathematics
- Loss type
- Unmigrated
- Stage
- Design, Evaluation
- Data
- Analytical / expert
- Effort
- Days
Steps
- Enumerate all interaction rules users must internalize to use the product correctly.
- Each rule is: 'when condition X, action Y produces result Z'.
- Count the rules and their exceptions.
- Each exception adds complexity.
- Compute a proxy KC: total rules + 2× (number of exceptions).
- Compare: if redesign reduces rule count by > 20%, predict measurable improvement in learnability.
- Target: < 15 rules for consumer products, < 30 for professional tools.
Sources
- Information Theory, Inference, and Learning Algorithms — David J.C. MacKay
- Rissanen, Modeling by shortest data description (Automatica, 1978)
- Payne & Green, Task-action grammars (Human-Computer Interaction, 1986)
Also known as
- Minimum description length of interaction grammars
- Compression-based consistency metrics
Categories
- Information Theory, Physics & Mathematics