jcp-efficiency: the outcome anchor is not ours alone; the signature is #319
2026-10-18
jcp-efficiency: the outcome anchor is not ours alone; the signature is
All checks were successful
standards workspaces build + test / which workspaces this change touches (pull_request) Successful in 22s
repository checks / lockfiles, tests in CI, workflow hygiene, spec and section references, public text (pull_request) Successful in 3m39s
standards workspaces build + test / jcp-rs (jcp-conformance, jcp/jcp-rs, -p jcp-energy --features nvml, jcp-rs, linux) (pull_request) Successful in 1m43s
standards workspaces build + test / joule-ui-rs (joule-ui/joule-ui-rs, joule-ui-rs, linux) (pull_request) Successful in 1m17s
standards workspaces build + test / joulehook-rs (joulehook/joulehook-rs, joulehook-rs, linux) (pull_request) Successful in 1m20s
standards workspaces build + test / sandbox-rs (sandbox-conformance, sandbox/sandbox-rs, sandbox-rs, linux) (pull_request) Successful in 2m26s
standards workspaces build + test / macOS energy backends (pull_request) Successful in 12m44s
standards workspaces build + test / jouledesk-rs (jouledesk/jouledesk-rs, jouledesk-rs, macos) (pull_request) Successful in 5m33s
standards workspaces build + test / jouleclaw-rs (-nodebug, jouleclaw/jouleclaw-rs, -p jouleclaw-energy --features rapl,nvml, jouleclaw-rs, linux-arm64, -C target-feature=+fp16) (pull_request) Successful in 3m59s
standards workspaces build + test / joule-code-rs (joule-code/joule-code-rs, joule-code-rs, linux-arm64, true) (pull_request) Successful in 3m10s
standards workspaces build + test / mesh-rs (mesh/mesh-rs, -p mesh-engine-gguf -p mesh-openai --features mesh-engine-gguf/gguf,mesh-openai/server, mesh-rs, linux-arm64) (pull_request) Successful in 2m3s
Three places claimed the outcome anchor as the thing nobody else has:
- `src/lib.rs` — "the four figures that matter, plus the one thing the
incumbents omit — the **outcome anchor**"
- `MAPPING.md` §0 — "seals all four **with the outcome anchor the rates
omit**"
- `jcp/README.md` — "with the one thing the rates omit: the **outcome
anchor**"
That is no longer true. Panigrahy, Tyagi, Abkin, Houngninou and
Mahapatra, *Energy per Successful Goal: Goal-Level Energy Accounting for
Agentic AI Systems* (arXiv:2605.22883, v1 2026-05-20, v2 2026-10-12),
defines EpG as "total workflow energy across all attempts, including
failures and retries, divided by successfully completed goals" — the
same denominator `uj_per_completed_unit()` computes, waste included —
with a measurement framework attributing RAPL and ARM counters to
workflow energy.
What survives the finding, and is now stated as the claim rather than
assumed inside a broader one: the paper is a *calculation and reporting*
convention. It describes measurement, attribution and a reproducibility
protocol; no part of it signs or attests the figure. So `efficiency/1`
adds not the anchor but the **artifact** — a signed, honest-provenance,
verifiable receipt carrying it. MAPPING.md already said exactly that in
its opening ("supplies the one thing that standard, and every deployed
disclosure regime, lacks: a signed, honest-provenance, verifiable
receipt carrying an outcome-anchored figure"), which is why that
sentence needed no change: it was scoped correctly and the three
downstream restatements were not.
MAPPING.md §0 now cites the prior art in the same ingest-not-compete
form §1 already uses for GSF SCI-for-AI, including the licence
constraint (arXiv non-exclusive — cited and briefly quoted, never
reproduced). Naming it is within the naming rule, which permits
standards, bodies, protocols, papers and open-source projects to be
named factually; `check-public-names.sh` scopes to `wai/` in any case.
Guards run locally before pushing, since this is public text:
comparative-prose clean (2307 files), no vanity metrics (277 Markdown
files), check-spec-refs 1613 citations all resolving.
`cargo test -p jcp-efficiency`: 54 + 6 pass.
Co-Authored-By: Claude Opus 5 <[email protected]>