Skip to content

Deduplicate recording proofs and clarify the memoryless invocation law - #64

Merged
MarcIlunga merged 1 commit into
mainfrom
codex/recording-proof-cleanup
Oct 29, 2026
Merged

MarcIlunga merged 1 commit into
mainfrom
codex/recording-proof-cleanup

Conversation

@MarcIlunga

Copy link
Copy Markdown
Collaborator

The probabilistic-automaton and program-attachment proofs repeated the recording invariant's support argument, and replyLaw named both an automaton step and a program invocation. Both proofs now reuse recording_reachable directly and share recording_init_reachable. The program's law explicitly names its memoryless inside replies; its definition and the public attachment statements keep their existing mathematical meaning.

Previous declaration Replacement
Program.replyLaw Program.memorylessReplyLaw
Program.replyLaw_inl Program.memorylessReplyLaw_inl
Program.replyLaw_succ_inr Program.memorylessReplyLaw_succ_inr

All repository consumers and documentation links are updated. The recursive reply law and the recording invariant are explained beside their definitions/proofs. Claude reviewed the comments with only file-reading/editing tools, with no shell, Git, external tools or hooks. The coordinator checked the resulting diff.

Validation: focused builds through PRG length extension and the systems axiom checks, standard Mathlib lint on every changed Lean file, the affected-module axiom audit, source checks and git diff --check. Classical openings in the ideal-system definitions were narrowed to their definitions to satisfy the standard linter.

Fixes #31.

@MarcIlunga
MarcIlunga merged commit 2b9487f into main Oct 29, 2026
4 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.

Follow-ups from the review of #22: a repeated proof block, and two meanings of replyLaw

1 participant