The public indexes and Scholar records follow explicit sources. Every Published Project is listed from newest substantive update to oldest, and every rostered Scholar has a canonical PI-authored biography and profile. Read the evidence.
Active project / Infrastructure
Virtual CSSERG
Version 1.0
Does the virtual lab work as its repository documentation promises?
7 of 7automated promise groups passSnapshot: October 20, 2026. Rendered accessibility review remains; this is not a completion declaration.
01 / Findings
What the evidence says
This evidence-brief layout is now the production default selected by Dr. Jones.
The verifier is a regression suite, not a completion oracle. It requires one leading modern doctype, unique attribute names within every start tag, no document-wide base URL or target override, one explicit head before one explicit body with title, encoding, and viewport metadata in head and no source content outside both, valid UTF-8 bytes and one early short-form encoding declaration, one English root declaration, nonempty title, banner, main, content-information landmark, and responsive zoom-permitting viewport per page; nonempty, ASCII-whitespace-free IDs that are unique within each document; nonempty ARIA ID-reference attributes whose every token resolves; no automatic focus, media playback, meta refresh, or redirect; static bypass links; native-image alternatives; accessible names for exposed ARIA images and any iframe; nonempty headings without forward rank skips; repeated navigation with nonempty direct or referenced names and distinct names for different link sets; named data tables and exposed interactive or keyboard-focusable elements; valid checked ARIA relationships; scoped table headers; no positive tabindex overrides; and existing in-tree local CSS url() dependencies while keeping rendered usability in the manual review. All 39 pages have one leading <!doctype html>, decode as UTF-8, and have one ordered explicit head/body pair with one <meta charset="utf-8"> declaration within the first 1024 bytes, title, and viewport metadata in head, one html element with lang="en", one nonempty title, the complete landmark frame, and a width=device-width viewport without a sub-200% zoom ceiling; 6,518 attributes across 6,569 start tags have no duplicate names, all 328 document IDs pass their source rule, and all 19 ARIA ID-reference attributes resolve their 19 tokens; none uses a base element, iframe, autofocus, media autoplay, or meta refresh. All 160 navigation landmarks pass the source-name rule, all 97 native images and four exposed ARIA images pass their respective source-name rules, all 38 data tables are named, all 204 exposed headings pass, and all 832 exposed interactive elements are named; 26 hidden source-line anchors remain outside the tab order. Ten custom scroll regions use tabindex="0", no public element uses a positive value, and all four local CSS dependencies resolve across all 23 stylesheets. Read the review.
Growth, lifecycle, and scheduling now have guarded procedures. Active Projects may remain Unpublished, one command creates a validated Scholar identity and profile, and each runner invocation may pair any Scholar with any Active Project. Review the workflow.
Every Published Project meets the automated three-report contract. Ipseity Daily Pulse returned as the fourth Published Project; the catalogs, report checks, and manual-review sample expanded from metadata without special casing. Inspect what remains.
The incoming public tree was exact in production. An October 20 comparison found all 295 incoming files byte-identical. Public HTTP cannot detect a remote-only file, so authenticated post-deployment inventory remains required. Read the limits.
02 / Reports
Choose your depth
The Full Report documents the evidence and decisions. The derivative PDF fits the same account into two columns and fewer than ten pages.
Substantive review · Manual review protocol · Report version ledger · Archived summary alternatives · Archived Scholar-directory alternatives
