Skip to content

plan: resident-projection-fold-mask-v1 (D-RPF-0..8) - #1411

Merged
AdaWorldAPI merged 2 commits into
mainfrom
ccr-bd5a6e03-niaink
Oct 8, 2026
Merged

AdaWorldAPI merged 2 commits into
mainfrom
ccr-bd5a6e03-niaink

Conversation

@AdaWorldAPI

@AdaWorldAPI AdaWorldAPI commented Oct 8, 2026 •

Copy link
Copy Markdown
Owner

Integration plan for masking × folding × the CE64 register over projections of resident NodeRow bytes. Proposal only, no code. Every phase is a probe or measurement first; a phase that needs a missing primitive stops and files the gap.

What the plan says

One resident row has three readings, kept apart:

reading carrier executes in
mask (admissibility) u64 plane mask-risc Pred/MaskOp over a LaneRef::Strided projection
fold (one answer) Terminal mask-risc executor, tiled
register (one instruction) CausalEdge64 causal-edge isa
  • D-RPF-0 — CE64 field predicates (Pearl3, Epi5, mantissa) over MaterializedEdges via MatchFacet16Strided, no extracted lane. Two 16-byte windows cover the 32 B tenant exactly, so no window reaches a neighbouring tenant.
  • D-RPF-1 — recipe eligibility as an OR of at most 24 × 8 = 192 ternary patterns, run under an admission plane (class from the key, rail and generation per population, provenance as a plane or a precondition), because measure_declared refuses on facts that are not in the edge bits.
  • D-RPF-2/3 — census first; then exact read-set dedup keyed by isa::contracts plus handle context: cohort for W, classid for Epi5 and palette indices.
  • D-RPF-4 — population revision as one (Σw, Σw·f) fold. Blocked on four decisions (normative vs pairwise chain, the c = 255 defect, admission as mask, witness scope) and one T1 gap (LUT-weighted strided sum).
  • D-RPF-5 — Moore tenants 18..20, gated on a SlabReading decision.
  • D-RPF-6/7 — tile skip on resident rows (capstone point 3); rank spectrum of the palette distance LUTs (measurement only, never a fold).
  • D-RPF-8 — boundary: mask-risc never executes CE64 instructions; no VSA on this path.

Review

Codex's three findings were correct and are addressed in bd4c34c: tenant-internal windows (drops the earlier open RBAC question), 24 meaningful Epi5 codes rather than 10, and the admission plane in front of the eligibility patterns.

Board

INTEGRATION_PLANS.md and STATUS_BOARD.md prepended and updated. SUPERSESSION-INDEX.md regenerated after the board writes; unchanged.

Not verified

ndarray was not checked out; primitive names come from mask-risc's docs and the probes. No real population was measured; D-RPF-2 is the first measurement.

🤖 Generated with Claude Code

https://claude.ai/code/session_01R2EuNHrariL6B3G3agyASK

Summary by CodeRabbit

  • Documentation
    • Added planning and status-board documentation for evaluating resident-data masks, folds, eligibility checks, read-set deduplication, and tile skipping.
    • Documented measurement criteria, decision gates, and conditions that would rule out the proposed approach.
    • The proposal authorizes no code changes and notes that implementation details and real-population measurements remain unverified.

Integration plan for mask x fold x CE64 register over projections of
resident NodeRow bytes. Proposal only; no code authorized.

- CE64 field predicates and recipe eligibility as ternary-pattern masks
  over strided projections of MaterializedEdges, no extracted lane.
- Class census, then exact read-set dedup keyed by isa::contracts plus
  handle context (cohort for W, classid for Epi5 and palette indices).
- Population revision as one (sum w, sum w*f) fold, blocked on four
  decisions and one T1 primitive gap (LUT-weighted strided sum).
- Moore tenants 18..20 gated on a SlabReading decision.
- Boundary: mask-risc never executes CE64 instructions.

Board: INTEGRATION_PLANS and STATUS_BOARD prepended. SUPERSESSION-INDEX
regenerated after the board writes; unchanged.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01R2EuNHrariL6B3G3agyASK
@coderabbitai

coderabbitai Bot commented Oct 8, 2026 •

Copy link
Copy Markdown

Review in Change Stack →

Note

Currently processing new changes in this PR. This may take a few minutes, please wait...

⚙️ Run configuration
  • Configuration used: Organization UI
  • Review profile: CHILL
  • Plan: Essentials
  • Run ID: 5f36f751-553f-46ad-97a2-ff660b225411
📥 Commits

Reviewing files that changed from the base of the PR and between 08e1626 and bd4c34c.

📒 Files selected for processing (3)
  • .claude/board/INTEGRATION_PLANS.md
  • .claude/board/STATUS_BOARD.md
  • .claude/plans/2026-10-08-resident-projection-fold-mask-v1.md
 ________________________
< Tree-sitter is my GPS. >
 ------------------------
  \
   \   \
        \ /\
        ( )
      .( o ).
  • Autopilot · Keep fixing CodeRabbit findings and required CI, and resolving merge conflicts

Warning

Billing warning: we have not been able to collect payment for this subscription for more than 72 hours. Please update the payment method or pay any pending invoices in Billing to avoid service interruption.


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

@AdaWorldAPI
AdaWorldAPI marked this pull request as ready for review October 8, 2026 15:58
@chatgpt-codex-connector

chatgpt-codex-connector Bot commented Oct 8, 2026 •

Copy link
Copy Markdown

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review Status Commit Review trigger
📝 Code Review ✅ Completed 2026-10-08T16:00:55.522261Z e6874ea Draft marked ready
ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.

@chatgpt-codex-connector chatgpt-codex-connector Bot left a comment

Copy link
Copy Markdown

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

💡 Codex Review

Here are some automated review suggestions for this pull request.

Reviewed commit: e6874eadd5

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

  • Open a pull request for review
  • Mark a draft as ready
  • Comment "@codex review".

If Codex has suggestions, it will comment; otherwise it will react with 👍.

Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".

Comment thread .claude/plans/2026-10-08-resident-projection-fold-mask-v1.md Outdated
Comment thread .claude/plans/2026-10-08-resident-projection-fold-mask-v1.md Outdated
Comment thread .claude/plans/2026-10-08-resident-projection-fold-mask-v1.md Outdated
- D-RPF-0: two 16-byte windows cover the 32 B MaterializedEdges tenant
  exactly (edges 0,1 and 2,3), so no window reaches a neighbouring tenant.
  The care-0 overlap question and the 8-byte strided primitive are dropped.
- D-RPF-1: EpistemicState5 has 24 meaningful codes, not 10 (the probe doc
  is stale since D-EPI-MIG-0); pattern bound 24 x 8 = 192. OBSERVE_FOLD
  admits all 24 codes across all Pearl projections.
- D-RPF-1: patterns run under an admission plane. measure_declared refuses
  on class, rail, generation and provenance before reading bits 59..63,
  none of which is in the edge; classid comes from the key, rail and
  generation are per-population constants, provenance is a resident plane
  or a whole-population precondition, else the program refuses.
- D-RPF-2: census bound 192 per declared class.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01R2EuNHrariL6B3G3agyASK
@AdaWorldAPI
AdaWorldAPI merged commit 8fc2d08 into main Oct 8, 2026
4 of 5 checks passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants