forked from ChelseaKR/mrf-honest
-
Notifications
You must be signed in to change notification settings - Fork 0
Expand file tree
/
Copy pathbaseline.json
More file actions
48 lines (48 loc) · 3.32 KB
/
Copy pathbaseline.json
File metadata and controls
48 lines (48 loc) · 3.32 KB
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
{
"meta": {
"commit": "116ded9368585f82de38cd7ae4421d320a140e9a",
"date": "2026-08-19",
"environment": "macOS 25.4.0 arm64, headless Chrome via lighthouse 12, site served by python -m http.server on loopback",
"tools": {
"lighthouse": "12",
"node": "26.7.0",
"python": "3.12"
},
"scope": "Every page the published cohorts render: the index, how-we-grade, one page per file row of the 2026-08-19 JSON cohort (17) and the 2026-08-19 CSV cohort (25), and 404.html. Every score below is the minimum observed across all of them, not an average and not the best page. Re-measured when the site grew from 6 file pages to 17, and re-audited on the publish path when it grew to 42: the counts here describe pages that are actually audited, so growing a cohort without the audit passing fails the deploy rather than shipping a figure measured over a smaller site.",
"pages_audited": 45,
"file_pages": 42,
"k6": "N/A. There is no server: the published artifact is static files on GitHub Pages, so there is no request-latency budget to measure. PERFORMANCE-STANDARD section 1's p95 rows are null below rather than silently absent."
},
"floors": {
"accessibility": 1.0,
"best-practices": 1.0,
"seo": 1.0,
"performance": 0.95
},
"metrics": {
"accessibility": 1.0,
"best-practices": 1.0,
"seo": 1.0,
"performance": 1.0,
"max_total_transfer_bytes": 52404,
"max_requests_per_page": 1,
"p95_ms": null,
"llm_first_token_ms": null,
"llm_full_response_ms": null,
"js_kb_gzip": 0
},
"direction": {
"accessibility": "higher_is_better",
"best-practices": "higher_is_better",
"seo": "higher_is_better",
"performance": "higher_is_better",
"max_total_transfer_bytes": "lower_is_better",
"max_requests_per_page": "lower_is_better",
"js_kb_gzip": "lower_is_better"
},
"notes": [
"The accessibility floor is 1.0, not the standard's 0.90. ACCESSIBILITY-STANDARD section 1 says a repository that declares a higher floor enforces the higher value. On a static site with no JavaScript there is no honest reason for a 0.99: every point below 100 is a specific element a specific person cannot use.",
"max_total_transfer_bytes rose from 12,197 to 34,938 when the cohort grew from 6 graded files to 17, and to 52,404 when the CSV cohort added its section: the heaviest page is still the index, now carrying two clearly scoped cohort sections, 42 file cards, and every recorded exclusion with its cross-reference. That is content, not weight -- the request count is still 1 and there is still no script, stylesheet, font, image, or third party on any page -- and it remains inside the 61,440-byte document budget in perf/resource-budget.json. The number is refreshed rather than ratcheted because it is a measurement of the artifact, not a target.",
"These were measured after two real defects were fixed. Before the fix the index scored 0.98 on accessibility (axe heading-order: the file cards' h3 followed h1 with no h2) and every file page carrying a warning finding scored 0.95 (axe color-contrast: the FINDINGS chip was 4.28:1 against its wash, below the 4.5:1 of SC 1.4.3). Both shipped publicly because the portfolio registry recorded this repository as having no HTML surface, so the accessibility gates were scoped out of a site that was already live."
]
}