METHOD FRAMEWORK · NO UNIVERSAL OPTIMUM
Context Budget: Allocate for the Task, Do Not Fill by Default
A long-scroll visual companion to the article. It summarizes the article’s evidence boundary and method; it does not extend the underlying result.
At a glance
Evidence in scope
- Keep binding requirements
- Retrieve only relevant evidence
- Remove stale/redundant context only after checks
Evidence boundary: The proportions in this conceptual map are not measured token shares or a recommended context percentage.
How to use the diagram
Map each context item to a requirement, source, task state, or output need; test changes against the same acceptance rubric.
What this visual explains: Treat context as a task-specific allocation that must preserve requirements and leave room for tools and the answer.
What it does not establish
No universal context size, token-saving percentage, or quality improvement is measured.