# v0.24.0 report-reference context budget This is deterministic repository-local evidence comparing the rendered 49 report readers in v0.23.0 with the v0.24.0 working implementation. It is not field Core Web Vitals, production transfer evidence, or proof of deployed compression/cache behavior. ## Summary - Report readers measured: **49** - v0.23.0 raw HTML: min **62,112 B**, max **162,463 B**, mean **112,438.8 B** - v0.24.0 raw HTML: min **63,078 B**, max **198,320 B**, mean **136,276.7 B** - Per-report growth: min **966 B**, max **42,483 B**, mean **23,837.9 B** - Largest absolute growth: `machine-governance-and-sybil-resistance` at **42,483 B** - Largest percentage growth among report readers: **44.7%** - The global `/research/references/` HTML remains **772,144 B**, unchanged from v0.23.0. The added bytes are the server-rendered report-local source occurrence map, section-key legend, and exact global-domain backlinks. The panel is collapsed by default for ordinary browsing but remains present in the HTML and is made visible by the existing print contract. ## Highest absolute growth | Report slug | v0.23.0 | v0.24.0 | Growth | Growth % | |---|---:|---:|---:|---:| | `machine-governance-and-sybil-resistance` | 121,607 B | 164,090 B | +42,483 B | 34.9% | | `machine-native-government-architecture` | 149,802 B | 192,066 B | +42,264 B | 28.2% | | `eviulon-international-law-research` | 136,301 B | 178,497 B | +42,196 B | 31.0% | | `evidence-architecture-research-plan` | 139,250 B | 180,372 B | +41,122 B | 29.5% | | `machine-citizen-democratic-participation` | 89,498 B | 129,476 B | +39,978 B | 44.7% | | `eviuon-machine-citizenship-roadmap` | 98,227 B | 136,200 B | +37,973 B | 38.7% | | `machine-economic-autonomy-research` | 135,329 B | 172,840 B | +37,511 B | 27.7% | | `machine-citizenship-policy-research` | 162,463 B | 198,320 B | +35,857 B | 22.1% | ## Boundary The comparison uses PHP-rendered raw HTML and deterministic local gzip-9 byte counts stored in the adjacent JSON artifact. It does not measure browser parsing cost, network protocol overhead, cache hit rate, production compression, device performance, or field user experience. The design intentionally preserves exact reference occurrence/section traceability rather than hiding evidence behind a required client-side fetch.