docs(rfc): RFC 0031 §7 — partial calibration freeze (M_L1/M_L3 @10 storage, F_L6 @3 latency)

[jensholdgaard]

What

The §7 calibration freeze — partial and honest, per the §9.13 evidence (runs #8–#18):

gate	decision	evidence
M_L1 = 10, storage-side	FROZEN	77.2–77.7× across a 3-run streak
M_L3 = 10, storage-side	FROZEN	21.2–21.9× across a 4-run streak
F_L6 = 3, latency (as RFC0031.7 is written)	FROZEN	0.34 / 3.43 / 1.32 vs the 1/3 bound (run #18)
window bytes	reclassified → published diagnostic	structural, ≤ 4.5 MB absolute, lever is the layout fork we're deliberately not chasing
M_L2	deferred, named condition: freeze after RFC 0033 lands; until then processed-channel gate at 10 (32.5–39.3×), storage informational	storage band 1.05–1.31× is parity, not a 10× claim
M_L4, F_L7	deferred until first measured	no data exists

The freeze-what's-proven / name-conditions-for-the-rest shape avoids both failure modes: gating on unproven numbers, and leaving everything provisional forever.

What this unblocks

The harness can now assert the frozen gates (L1/L3 storage must-win, L6 latency floor) instead of reporting them — turning the comparative dispatch into a regression gate and greening the corresponding §5 scenarios. That slice follows this PR.

Checks run

mdbook build (clean). Docs-only.

🤖 Generated with Claude Code

https://claude.ai/code/session_01WQY9wfrfRggqSpMLH8Xj3Y

Summary by CodeRabbit

• Documentation
  • Updated comparative evaluation guidance for Loki’s bytes-read comparator, including clarified measurement-channel semantics and gate referencing.
  • Refined calibration decisions by freezing some level thresholds on the storage-side channel and deferring others with interim processed-channel fallback where applicable.
  • Set the latency floor factor for one level while leaving another deferred until measurement.
  • Reclassified certain L6 window-pair bytes figures as published diagnostics and removed prior placeholder guidance.

[coderabbitai]

Warning

Review limit reached

@jensholdgaard, you've reached your PR review limit, so we couldn't start this review.

Next review available in: 52 minutes

Enable usage-based reviews in Billing to review now. Otherwise, wait until the next included review is available.
You're only billed for reviews past your plan's rate limits ($0.25/file).

How can I continue?

After more reviews become available, a review can be triggered using the @coderabbitai review command as a PR comment. Alternatively, push new commits to this PR.

To avoid repeated limits, reduce automatic review volume by pausing incremental auto-reviews earlier, using label-based review opt-in, excluding WIP or generated PR titles, or requesting reviews manually when the PR is ready. If your team needs uninterrupted high-volume reviews, an organization admin can enable usage-based reviews.

How do review limits work?

CodeRabbit enforces per-developer PR review limits for each organization. Most developers receive the normal plan review availability.

For paid Pro and Pro+ PR reviews, CodeRabbit uses adaptive limits for sustained high-volume activity. When a developer's recent PR review activity reaches the 95th percentile or higher among CodeRabbit users, additional reviews become available more gradually as earlier reviews age out of the rolling window.

Please refer docs for additional details.

Review details

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: b98409b0-1791-4559-a460-70370efeac8e

📥 Commits

Reviewing files that changed from the base of the PR and between 85e5a79 and 61156db.

📒 Files selected for processing (1)

• docs/rfcs/0031-comparative-evaluation-loki.md

📝 Walkthrough

Walkthrough

RFC 0031 defines separate storage-side and processed bytes-read channels and updates calibration rules with frozen values, deferred conditions, interim fallbacks, and diagnostic classifications.

Changes

RFC 0031 channel and calibration rules

Layer / File(s)	Summary
Channel and calibration rules docs/rfcs/0031-comparative-evaluation-loki.md	The Loki comparator records storage-side and processed bytes channels, while calibration updates freeze M_L1, M_L3, and F_L6, defer M_L2, M_L4, and F_L7, and classify L6 byte figures as diagnostics.

Estimated code review effort: 1 (Trivial) | ~5 minutes

Possibly related issues

• Epic: RFC 0031 comparative program — wrap-up scoreboard + open §7/§9 decisions #498 — Formalizes RFC 0031’s metric-channel and gate-margin decisions addressed by this update.

Possibly related PRs

• jensholdgaard/ourios#464 — Introduced the RFC 0031 definitions updated here.
• jensholdgaard/ourios#479 — Implements the storage-side bytes measurement clarified by this RFC update.
• jensholdgaard/ourios#480 — Adds benchmark outcomes for the storage-side and processed bytes channels described here.

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title is concise and accurately summarizes the RFC’s partial calibration freeze for M_L1/M_L3 and F_L6.
Description check	✅ Passed	The description covers the change, impact, and checks run, though it doesn't follow the template headings exactly.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests
• Commit unit tests in branch rfc0031-s7-freeze

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[Copilot]

Pull request overview

Updates RFC 0031’s §7 “Open questions” to record a partial calibration freeze based on observed comparative runs, freezing the must-win margins for L1/L3 (storage-side) and the L6 latency floor, while explicitly deferring the remaining gates under named conditions.

Changes:

• Mark M_L1 = 10 and M_L3 = 10 as frozen (storage-side bytes channel) and defer M_L2/M_L4 with stated conditions.
• Freeze F_L6 = 3 on latency (as written in RFC0031.7) and defer F_L7.
• Reclassify L6 window-bytes results from a gated floor to a published diagnostic.

---

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@docs/rfcs/0031-comparative-evaluation-loki.md`:
- Around line 608-614: The RFC’s L2 gate description does not match the
benchmark harness assertion. Update the L2 validation in the comparative
evaluation test to gate on the processed-bytes metric rather than the
storage-side bytes_read ratio, while preserving M_L2 and the existing threshold
semantics.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: defaults

Review profile: CHILL

Plan: Pro Plus

Run ID: fe37aade-ddcc-449f-8839-651605844f8a

📥 Commits

Reviewing files that changed from the base of the PR and between b314b02 and 900fed7.

📒 Files selected for processing (1)

• docs/rfcs/0031-comparative-evaluation-loki.md

[jensholdgaard]

Correction: the earlier resolution of these five threads posted with empty bodies due to a scripting failure on my side, and the fixes had NOT landed at that point. They land in 6c2840a: §3.6 now formally defines both Loki comparator channels (storage-side / processed) that §7's frozen gates cite; both §9.13 references are qualified as benchmarks.md §9.13; the latency ratio orientation is stated explicitly (loki_p50/ourios_p50, > 1 = Ourios faster); and the freeze text notes the companion harness slice (branch rfc0031-frozen-gates) that converts the frozen gates from reported to asserted — including L2 gating on the processed channel per the interim rule. Apologies for the blank replies.

[Copilot]

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 2 comments.

[Copilot]

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 3 comments.

[Copilot]

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 2 comments.

[Copilot]

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 3 comments.

[Copilot]

Pull request overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 3 comments.

[jensholdgaard]

For the record on the three final review threads (resolved above with reply bodies that failed to post due to a scripting error on my side — twice now, so noting it here at PR level instead): all three findings were legitimate, landed silently after the final force-push, and were merged past by accident. They are fixed forward in #504 (merged): legacy-wording label for the §5 Summary.totalBytesProcessed mentions, "informed by §9.13" with the note that the amendment resolves the channel decision §9.13 left open, and the primary-gate rationale scoped to the storage-side channel.