code wiki / _hdl_build / nx_instrument_status.nx
nx_instrument_status.nx source
↩ module page · 147 lines · 13845 B
1// nx_instrument_status.nx -- GENERATOR: the living instrument dashboard (operator 2026-07-10 "make sure all our
2// gains are live and wired" + the standing "compare = living as-of-today"). Emits knowledge/instrument.html: a
3// self-contained, designed status page for the SOTA-evaluation instrument built 2026-07-09..10 -- integrity,
4// eval-methodology vs the field, the pipeline, the living product bars (repo-health verified), and the converted
5// censuses. Numbers are TRANSCRIBED FROM the gates that compute them (cited per row); the page carries its
6// regen command + date so it never silently goes stale (coach C4 discipline). Published via nx_instrument_ship
7// -> nx_aw_push -> https://nishifamily.com/instrument. NOT a gated prefix = serves public. license_tier: ORIGINAL
8import "nx_syscalls.nx"
9
10func hw(fd: i64, s: *u8) -> i64 { var n: i64 = 0; while s[n] != (0 as u8) { n = n + 1 } sys_write(fd, s, n); return 0 }
11func hn(fd: i64, v: i64) -> i64 {
12 let bb: *u8 = sys_mmap(28)
13 var m: i64 = v
14 if m < 0 { hw(fd, "-" as *u8); m = 0 - m }
15 let t: *u8 = sys_mmap(28)
16 var k: i64 = 0
17 if m == 0 { t[0] = 48 as u8; k = 1 }
18 while m > 0 { t[k] = (48 + (m % 10)) as u8; m = m / 10; k = k + 1 }
19 var i: i64 = 0
20 while i < k { bb[i] = t[k - 1 - i]; i = i + 1 }
21 sys_write(fd, bb, k)
22 return 0
23}
24// a stat card: label, big value string, sub
25func card(fd: i64, label: *u8, val: *u8, sub: *u8) -> i64 {
26 hw(fd, "<div class='card'><div class='cl'>" as *u8); hw(fd, label)
27 hw(fd, "</div><div class='cv'>" as *u8); hw(fd, val)
28 hw(fd, "</div><div class='cs'>" as *u8); hw(fd, sub); hw(fd, "</div></div>" as *u8)
29 return 0
30}
31// an axis row: name, us, max, verdict-class (0 behind /1 parity), evidence
32func axrow(fd: i64, nm: *u8, us: i64, mx: i64, ev: *u8) -> i64 {
33 hw(fd, "<tr><td class='ax'>" as *u8); hw(fd, nm)
34 hw(fd, "</td><td class='sc'>" as *u8); hn(fd, us); hw(fd, "/" as *u8); hn(fd, mx)
35 if us >= mx { hw(fd, "</td><td><span class='v par'>PARITY</span></td><td class='ev'>" as *u8) }
36 else { hw(fd, "</td><td><span class='v beh'>BEHIND</span></td><td class='ev'>" as *u8) }
37 hw(fd, ev); hw(fd, "</td></tr>" as *u8)
38 return 0
39}
40// a product-bar row
41func barrow(fd: i64, nm: *u8, commits: *u8, ci: *u8, health: *u8, note: *u8) -> i64 {
42 hw(fd, "<tr><td class='pn'>" as *u8); hw(fd, nm)
43 hw(fd, "</td><td>" as *u8); hw(fd, commits)
44 hw(fd, "</td><td>" as *u8); hw(fd, ci)
45 hw(fd, "</td><td>" as *u8); hw(fd, health)
46 hw(fd, "</td><td class='ev'>" as *u8); hw(fd, note); hw(fd, "</td></tr>" as *u8)
47 return 0
48}
49// a converted-census row
50func cvrow(fd: i64, nm: *u8, score: *u8, note: *u8) -> i64 {
51 hw(fd, "<tr><td class='pn'>" as *u8); hw(fd, nm)
52 hw(fd, "</td><td><span class='ok'>COACH-OK</span></td><td class='sc'>" as *u8); hw(fd, score)
53 hw(fd, "</td><td class='ev'>" as *u8); hw(fd, note); hw(fd, "</td></tr>" as *u8)
54 return 0
55}
56
57func main() -> i64 {
58 let fd: i64 = sys_openat_wr("knowledge/instrument.html\x00" as *u8, 420)
59 if fd < 0 { hw(1, "cannot open instrument.html\n" as *u8); return 1 }
60
61 hw(fd, "<!doctype html><html lang='en'><head><meta charset='utf-8'><meta name='viewport' content='width=device-width,initial-scale=1'><title>Nishi Instrument -- living SOTA-evaluation dashboard</title><style>" as *u8)
62 hw(fd, ":root{--bg:#0b0d12;--pan:#12151d;--pan2:#161b28;--bd:#232c3f;--fg:#cdd6e4;--dim:#8fa3bf;--acc:#7fd4ff;--par:#57d08a;--beh:#e6a86a;--ok:#57d08a}" as *u8)
63 hw(fd, "*{box-sizing:border-box}html,body{margin:0;background:var(--bg);color:var(--fg);font-family:'Segoe UI',system-ui,sans-serif;line-height:1.5}" as *u8)
64 hw(fd, "a{color:var(--acc);text-decoration:none}a:hover{text-decoration:underline}" as *u8)
65 hw(fd, ".wrap{max-width:1000px;margin:0 auto;padding:0 20px 80px}" as *u8)
66 hw(fd, "header{border-bottom:1px solid var(--bd);background:linear-gradient(180deg,#12151d,#0b0d12);padding:34px 20px 26px;text-align:center}" as *u8)
67 hw(fd, ".brand{font-size:13px;letter-spacing:.28em;color:var(--acc);font-weight:600}" as *u8)
68 hw(fd, "h1{margin:8px 0 6px;font-size:26px;font-weight:600;letter-spacing:-.01em}" as *u8)
69 hw(fd, ".tag{color:var(--dim);font-size:14px;max-width:640px;margin:0 auto}" as *u8)
70 hw(fd, ".date{margin-top:12px;font-size:11px;color:#5f7397;font-family:Consolas,monospace}" as *u8)
71 hw(fd, "h2{font-size:12px;letter-spacing:.14em;color:var(--acc);text-transform:uppercase;margin:38px 0 14px;border-bottom:1px solid var(--bd);padding-bottom:8px}" as *u8)
72 hw(fd, ".cards{display:grid;grid-template-columns:repeat(auto-fit,minmax(180px,1fr));gap:14px}" as *u8)
73 hw(fd, ".card{background:var(--pan);border:1px solid var(--bd);border-radius:12px;padding:16px 18px}" as *u8)
74 hw(fd, ".cl{font-size:11px;letter-spacing:.08em;color:var(--dim);text-transform:uppercase}" as *u8)
75 hw(fd, ".cv{font-size:30px;font-weight:700;color:#eaf2ff;margin:6px 0 2px;font-variant-numeric:tabular-nums}" as *u8)
76 hw(fd, ".cs{font-size:12px;color:var(--dim)}" as *u8)
77 hw(fd, "table{width:100%;border-collapse:collapse;font-size:13px;background:var(--pan);border:1px solid var(--bd);border-radius:12px;overflow:hidden}" as *u8)
78 hw(fd, "th{text-align:left;font-size:10px;letter-spacing:.1em;text-transform:uppercase;color:var(--dim);padding:10px 12px;background:var(--pan2);border-bottom:1px solid var(--bd)}" as *u8)
79 hw(fd, "td{padding:9px 12px;border-bottom:1px solid #1a2130;vertical-align:top}" as *u8)
80 hw(fd, "tr:last-child td{border-bottom:none}.ax,.pn{color:#e2eaf6;font-weight:500;white-space:nowrap}" as *u8)
81 hw(fd, ".sc{font-family:Consolas,monospace;color:#bcd;white-space:nowrap}.ev{color:var(--dim);font-size:12px}" as *u8)
82 hw(fd, ".v{font-size:10px;font-weight:700;padding:2px 7px;border-radius:6px;letter-spacing:.06em}" as *u8)
83 hw(fd, ".par{background:#153524;color:var(--par)}.beh{background:#3a2a17;color:var(--beh)}" as *u8)
84 hw(fd, ".ok{font-size:10px;font-weight:700;color:var(--ok);background:#153524;padding:2px 7px;border-radius:6px}" as *u8)
85 hw(fd, ".flow{display:flex;flex-wrap:wrap;gap:8px;align-items:center;font-size:12px;color:var(--dim)}" as *u8)
86 hw(fd, ".step{background:var(--pan2);border:1px solid var(--bd);border-radius:20px;padding:5px 13px;color:#cfe3ff}" as *u8)
87 hw(fd, ".arrow{color:#3d5a8f}p.note{color:var(--dim);font-size:13px}.hi{color:#eaf2ff}" as *u8)
88 hw(fd, "footer{margin-top:44px;padding-top:18px;border-top:1px solid var(--bd);color:#5f7397;font-size:11px;font-family:Consolas,monospace}" as *u8)
89 hw(fd, "@media (prefers-color-scheme:light){:root{--bg:#f4f6fa;--pan:#fff;--pan2:#eef2f8;--bd:#d5deeb;--fg:#1b2430;--dim:#5f7088;--acc:#1667b8}}" as *u8)
90 hw(fd, "</style></head><body>" as *u8)
91
92 hw(fd, "<header><div class='brand'>NISHI INSTRUMENT</div><h1>Living SOTA-evaluation dashboard</h1>" as *u8)
93 hw(fd, "<div class='tag'>The sovereign tooling that evaluates any program, product, app, capability, ecosystem or suite from its git, docs, schematics and papers — against the current field, at zero token cost, with drift and false-SOTA actively liar-killed.</div>" as *u8)
94 hw(fd, "<div class='date'>generated 2026-07-10 · regen: nx_instrument_status → nx_instrument_ship → nx_aw_push</div></header><div class='wrap'>" as *u8)
95
96 hw(fd, "<h2>Instrument state</h2><div class='cards'>" as *u8)
97 card(fd, "Eval-methodology" as *u8, "625‰" as *u8, "[15/24] vs the 2025-26 field (Scorecard, CHAOSS, Lighthouse, SWE-bench, LLM-judge, G-Eval, SPDX)" as *u8)
98 card(fd, "Census integrity" as *u8, "7 / 87" as *u8, "product-surface censuses now carry honest UI/product axes (was 3; climbing per conversion)" as *u8)
99 card(fd, "Audit sweep" as *u8, "162" as *u8, "censuses scanned; 83 flagged for missing U-axes = the conversion queue" as *u8)
100 card(fd, "Research cost" as *u8, "0" as *u8, "LLM agents per query (field bar: 107) — sovereign TLS banking, keyless" as *u8)
101 hw(fd, "</div>" as *u8)
102
103 hw(fd, "<h2>How it evaluates — the pipeline</h2><div class='flow'>" as *u8)
104 hw(fd, "<span class='step'>bank the whole product</span><span class='arrow'>→</span>" as *u8)
105 hw(fd, "<span class='step'>discover axes from the field</span><span class='arrow'>→</span>" as *u8)
106 hw(fd, "<span class='step'>verify bars are living</span><span class='arrow'>→</span>" as *u8)
107 hw(fd, "<span class='step'>score us vs the bar</span><span class='arrow'>→</span>" as *u8)
108 hw(fd, "<span class='step'>coach-gate the claim</span></div>" as *u8)
109 hw(fd, "<p class='note'>Open-source products are the full-stack ground truth — their repos expose UI, backend, infra, networking and mobile in one place. Axes are derived from the banked corpora (cross-product corroboration), not from what we happened to build. A claim of parity or exceed <span class='hi'>cannot be made without a paired gate / live-URL evidence token</span> — the coach refuses it otherwise.</p>" as *u8)
110
111 hw(fd, "<h2>Evaluating the evaluator — vs the 2025-26 field</h2><table><tr><th>axis</th><th>us</th><th>verdict</th><th>evidence</th></tr>" as *u8)
112 axrow(fd, "Papers / frontier banking" as *u8, 2, 2, "keyless OpenAlex/arXiv pipeline, dated banks across 5 domains" as *u8)
113 axrow(fd, "Benchmark harness discipline" as *u8, 2, 2, "gates + liar-kills + neg-controls + determinism floors (SWE-bench-class)" as *u8)
114 axrow(fd, "Meta-evaluation (self-audit)" as *u8, 2, 2, "this dashboard + the audit + the coach + a tracked integrity ledger row" as *u8)
115 axrow(fd, "Repo security posture (Scorecard)" as *u8, 1, 2, "CI presence + maintained-ness from banked GitHub evidence" as *u8)
116 axrow(fd, "Community health (CHAOSS)" as *u8, 1, 2, "contributor base + community health_percentage from banked profiles" as *u8)
117 axrow(fd, "Doc/feature extraction" as *u8, 1, 2, "cross-product corroborated term discovery (782 terms); synonym grouping = next" as *u8)
118 axrow(fd, "Automated product audits (Lighthouse)" as *u8, 1, 2, "ui-judge / contrast / exceed graders exist; census-wide wiring = next" as *u8)
119 axrow(fd, "Model-as-judge (G-Eval)" as *u8, 1, 2, "domain judges gated; sovereign LLM-judge buildable (no-float model generates)" as *u8)
120 axrow(fd, "Living refresh / staleness" as *u8, 1, 2, "dated banks + current-quarter freshness check; scheduled cadence = next" as *u8)
121 axrow(fd, "SBOM / composition (SPDX)" as *u8, 1, 2, "stack-audit on our own 0-dependency stack; interop analysis = next" as *u8)
122 axrow(fd, "Hardware / schematic eval" as *u8, 1, 2, "gate-emit ALU/decoder verification; ERC/DRC breadth = next" as *u8)
123 axrow(fd, "Repo structure / layer analysis" as *u8, 1, 2, "layer fingerprint from repo tree; CI/release-cadence depth = next" as *u8)
124 hw(fd, "</table>" as *u8)
125
126 hw(fd, "<h2>The bars are living — verified today</h2><table><tr><th>product</th><th>recent commits</th><th>CI</th><th>community health</th><th>role</th></tr>" as *u8)
127 barrow(fd, "Godot 4.x" as *u8, "29 / 30" as *u8, "yes" as *u8, "100%" as *u8, "game-engine bar: editor / rendering / authoring" as *u8)
128 barrow(fd, "Luanti" as *u8, "30 / 30" as *u8, "yes" as *u8, "87%" as *u8, "voxel-sandbox bar (Minecraft-class OSS); ships mobile (android)" as *u8)
129 barrow(fd, "Veloren" as *u8, "26 / 30" as *u8, "yes" as *u8, "50%" as *u8, "3D multiplayer bar: shaders / world / netcode" as *u8)
130 hw(fd, "</table><p class='note'>Comparisons are against products proven maintained <span class='hi'>this week</span> — not a stale snapshot. Repo health is a Scorecard/CHAOSS-class read from banked evidence; an unbanked repo scores NO-EVIDENCE, never “healthy”.</p>" as *u8)
131
132 hw(fd, "<h2>Converted censuses — drift killed, truth stated</h2><table><tr><th>census</th><th>coach</th><th>product truth</th><th>the honest read</th></tr>" as *u8)
133 cvrow(fd, "cadtwin (CAD twin)" as *u8, "578‰" as *u8, "STEP interop + metrology strong; viewer visuals climbing (live)" as *u8)
134 cvrow(fd, "game s-class" as *u8, "357‰" as *u8, "foundation-strong, product-weak; U4 polish = 0 (‘toy level lives here’)" as *u8)
135 cvrow(fd, "game asset-gen" as *u8, "250‰" as *u8, "primitives exceed structurally; outputs toy-level vs shipped art" as *u8)
136 cvrow(fd, "game genre" as *u8, "125‰" as *u8, "coverage-strong; an emitted genre is a mechanic demo until juice/UI/audio land" as *u8)
137 cvrow(fd, "minecraft bench" as *u8, "structured" as *u8, "honestly BEHIND on graphics+gameplay; AHEAD only on sovereignty/P2P" as *u8)
138 hw(fd, "</table><p class='note'>Every score above <span class='hi'>dropped</span> when the UI/product axes landed — that fall is the false-SOTA signal being killed. A census only earns its claims once the coach confirms every claim carries its proof.</p>" as *u8)
139
140 hw(fd, "<h2>Live example</h2><p class='note'>A product this instrument both evaluated and drove to a live surface: <a href='/explodelab'>the CAD explode-view</a> — real STEP geometry (53/53 faces, micron-verified), dual-dimensioned mm+thou BOM, lit and anti-aliased, in sovereign integer wasm. Its own viewer-fidelity axes are scored honestly on the cadtwin row above.</p>" as *u8)
141
142 hw(fd, "<footer>Sovereign — hardware-up NishiLang, zero third-party dependencies, zero token cost. Every number here is transcribed from a gate that computes it (nx_evalsota_census / nx_census_uiaudit / nx_repo_health_gate / nx_ws_coach / nx_sota_gap_ledger) and re-verified on regeneration. No wave; measured only.</footer>" as *u8)
143 hw(fd, "</div></body></html>\n" as *u8)
144 sys_close(fd)
145 hw(1, "nx_instrument_status: wrote knowledge/instrument.html\n" as *u8)
146 return 0
147}