Clear public files matter.
This principle remains part of the current Digital Karma direction.
Changelog
A public timeline of what changed, when it changed, and which version is current.
specs/federation-spec-v8.md
This principle remains part of the current Digital Karma direction.
This principle remains part of the current Digital Karma direction.
This principle remains part of the current Digital Karma direction.
This principle remains part of the current Digital Karma direction.
This principle remains part of the current Digital Karma direction.
The changelog explains what changed. The website audit shows how those changes affect a live public implementation right now.
A shorter public-facing changelog lives here. The fuller article version is also available in the resource library.
Endpoint files could pass while the HTML pages exposed no Dataset graph. v8.2 closed that gap by requiring every canonical sitemap page to render the site's DataCatalog and complete catalog Dataset inventory, plus a generated /ai/rendered-schema.json receipt from a passing full-sitemap crawl. The manifest, health endpoint, nightly runner, inventory audit, and federation audit now carry that proof through the release process.
The same builder that writes llm.json now writes /llms.txt in the llmstxt.org format on every site, leaving any site-owned llms.txt alone; 160 sites serve one. Warehouse logs for 2026-08-18 to 2026-09-22 showed classified AI crawlers requesting llm.json 205 times against 115 manifest requests, and /llms.txt requested 2,116 times with roughly 960 answered 404. New knowledge-base article explains the decision and its limits. The download kit now ships tools/build-llm-json.php, and the nightly conformance audit checks llms.txt on every site.
llm.json returned as a generated endpoint because AI crawlers kept requesting it after v8.0 retired it. One builder (build-llm-json.php) now derives /llm.json and /ai/llm.json from each site's manifest nightly. The same release gave every site its full constellation peer list (most had 3 or 4 of 5 to 36 real siblings) and resolved sibling URLs to each site's declared origin. 1,085 endpoint files across 155 sites (plus midlandinfosys.net on its own account) carry protocol_version 8.1, checked nightly by audit-federation.php. See specs/federation-spec-v8.md.
Shared generation engine, protocol/lifecycle stamping, and single-sourced constellation membership rolled out to 90 of ~99 properties. Real drift found and fixed along the way on six sites (NaturalHerbLibrary.com, AgeBetterToday.com, RealSEOLife.com, krisada.com, and two AS400 IBM System sites), most with related_sites missing the majority of their real constellation siblings. See specs/federation-spec-v8.md for the full spec.
The prior 'v8.0 Constellation terminology rollout' entry below described a Stars/Alignments/Signals renaming that was never actually implemented anywhere in the portfolio. Marked superseded rather than deleted, so the historical record shows a direction that was considered and dropped, not silently erased.
Krisada.com formally designated as the Galaxy Historian and Librarian of the Digital Karma Constellation network. As the oldest domain in the portfolio with the deepest SEO lineage, it preserves institutional memory, chronicles the federation's history across versions, and anchors authority signals for the wider galaxy. Documentation across the federation updated to reflect this role.
A public wording shift was proposed, but it never shipped. The real v8.0 release was the later architecture update on 2026-07-29.
Aligned the public standards, developer docs, registry copy, and machine-readable indexes so the live site now consistently reflects the v7.1 root discovery rules and current constellation version.
Added a public validator and upgrade planner that can audit remote public sites against the latest standard, generate action plans, and export upgrade packages.
Published v7.1-DigitalKarma.md as the portable installer brief and aligned README and authority docs around the new root placement rules.
Clarified that robots.txt and sitemap.xml belong at the root and should not be relocated into /ai/.
Introduced the 6 required artifacts, formal compliance tiers, constellation mapping, and the modern Digital Karma badge naming.
Added /entities/ JSON-LD conventions and /ai/diagnostics.json as part of the higher-tier maturity path.
Four core AI endpoints formed the original structured baseline before the dedicated constellation layer existed.