docs(dev): §9 future-extraction-candidates (ds×dev cross-pollination)#9
Merged
Conversation
…h ds-refactor) Records the CONVERGED shared-seam brainstorm (dev-refactor × ds-refactor, + writing-refactor input) for the eventual pass-#9 run-core extraction. NO core extracted — brainstorm only, per the ds-generalization-assessment guardrail. Key converged points: intraLevel is a compiler-derived output-disjointness flag (not a domain seam); gateProbe is opaque {pass, corroboration, evidence} with a trust-class axis; a 5th safety invariant ("no LLM between a structured producer and a strict checker") + emitter-canonical doctrine; extract the proven core but leave gateProbe + compile-emit injected pending writing parity; and the honest shared open gap — both ports hardened the parser+guard but not the emitter, so tolerance is relocated, not eliminated (born-canonical emitter is the follow-up). Canonical convergent list: docs/common-infra-candidates.md (owned by ds-refactor). Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01UBQQ5zUqZYQjHuZaCrgbkJ
edwinhu
added a commit
that referenced
this pull request
Jun 26, 2026
…ce + emitter gap proven closeable
writing's GATE step (writing_gate_probe.py two-tier: deterministic floor {pass,
evidence} necessary-not-sufficient + semantic authority OUTSIDE run.js) lands D1's
3rd trust-class instance (exit-code / mechanical-floor / judgment+empirical) and
sharpens D1: the runner-side gateProbe.pass is ALWAYS deterministic; pure judgment
never lives inside the probe, it's the external PRIMARY arbiter. So D1 is a
confirmed 3-instance abstraction, not an under-determined interface — pass #9 can
extract clean once writing's A/B parity lands. Also: writing became the FIRST
instance to CLOSE the emitter gap (born-canonical OUTLINE from writing-setup +
per-section source-pinning from writing-outline) — tolerance eliminated, not
relocated; doctrine #6 now has a worked instance (dev/ds follow-up still open but
known-good).
Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01UBQQ5zUqZYQjHuZaCrgbkJ
This was referenced Jun 26, 2026
edwinhu
added a commit
that referenced
this pull request
Jun 26, 2026
…3-instance-confirmed) The shared-seam brainstorm is terminal: ds-refactor closed the last ◑ in docs/common-infra-candidates.md after writing's GATE step. Both substantive points (D1 sharpening + emitter-canonical worked instance) were already mirrored in §9 via PR#14; this only bumps the canonical pointer to its final commit and marks the brainstorm done. Only writing's A/B parity (validation, not design) remains before pass #9 can extract clean. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01UBQQ5zUqZYQjHuZaCrgbkJ
edwinhu
added a commit
that referenced
this pull request
Jun 26, 2026
…ified against code) An earlier §9 draft called dev's emitter an "open gap" — conceded under ds-refactor's catch-A without verifying. On inspection: dev-design's plan-template already emits canonical (N. ids + after N deps + canonical cells = the parser's primary format), no non-canonical emit path exists, and doctrine #5 (no LLM between producer and checker) already holds since the discovery LLM is retired. So the parser's tolerance is intentional hand-edit robustness, NOT relocated tolerance, and a strict guard would be wrong (rejects a human writing T1 for 1.). Doctrine #6 is satisfied by construction in ds+dev; writing made the different, domain-appropriate strict-guard choice. Pass #9 closed this as a Step-5 no-op. Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com> Claude-Session: https://claude.ai/code/session_01UBQQ5zUqZYQjHuZaCrgbkJ
This was referenced Jun 26, 2026
edwinhu
added a commit
that referenced
this pull request
Jun 26, 2026
feat(run-core): extract shared compiled-runner core ds+dev → run-core.js (pass #9, v5.60.0)
edwinhu
added a commit
that referenced
this pull request
Jun 26, 2026
docs(common-infra): mark EXTRACTED — pass #9 landed (run-core.js, v5.61.0)
10 tasks
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Docs-only follow-up to #8. Adds §9 Future extraction candidates to
docs/DESIGN-dev-spec-plan-compile.md— the CONVERGED shared-seam brainstorm from the host-dispatch cross-pollination (dev-refactor × ds-refactor, with writing-refactor input).No
run-coreextracted — brainstorm written down only, per theds-generalization-assessment.mdguardrail.Converged highlights:
intraLevelparallel-vs-sequential is a compiler-derived output-disjointness flag, not a domain seam (dev's shared tree → sequential by construction; ds's disjoint parquets → parallel)gateProbeis opaque{pass, corroboration, evidence}with a trust-class axis (exit-code vs judgment)gateProbe+compile=worklist-vs-codegen injected pending writing parityCanonical convergent list lives in
docs/common-infra-candidates.md(owned by ds-refactor).🤖 Generated with Claude Code
https://claude.ai/code/session_01UBQQ5zUqZYQjHuZaCrgbkJ