Skip to content

docs(research): archive session lifecycle experiment - #948

Open
benvinegar wants to merge 1 commit into
mainfrom
docs/archive-session-lifecycle-research
Open

docs(research): archive session lifecycle experiment#948
benvinegar wants to merge 1 commit into
mainfrom
docs/archive-session-lifecycle-research

Conversation

@benvinegar

Copy link
Copy Markdown
Member

Summary

  • add a repository-level research/ convention for bounded engineering experiments
  • archive the discarded session-broker Effect lifecycle experiment with its plan, final report, and selected control/treatment evidence
  • retain reconstructable patches, raw benchmark results, ownership inventories, source identities, validation manifests, and checksums without merging the treatment or its dependency
  • mark archival artifacts as generated and exclude byte-exact evidence JSON from formatting

Why

The experiment produced a useful null result and several actionable plain-TypeScript lifecycle lessons. Keeping only a local worktree or ignored tmp/ evidence would make those findings easy to lose and invite future maintainers to repeat the same investigation. This archive makes the discarded status, evidence boundary, costs, defects, and reconsideration criteria durable and reviewable.

Non-goals

  • no Effect dependency or treatment implementation enters production
  • no session-broker behavior, protocol, authentication, or public API changes
  • no claim that local hashes prove authorship, trusted time, or universal benchmark results

Validation

  • sha256sum -c research/session-broker-effect-lifecycle/artifacts/MANIFEST.sha256
  • bun run format:check
  • bun run lint
  • bun run typecheck
  • bun run deps:check
  • bun run changeset:status
  • independent archive and final staged-diff reviews found no blocker or high-severity issues

No runtime test suite was rerun for this PR because it changes only documentation, research artifacts, formatting exclusions, Git attributes, and an empty Changeset. The archived report separately records the experiment's original runtime validation.

Known limitations

  • selected structured evidence retains local username, CPU, and kernel provenance but no credentials or live secrets
  • bulky generated bundles and repetitive raw logs are omitted; the structured result files and original manifests retain their measurements and identities
  • reconstruction patches are marked binary in Git to preserve their exact whitespace and avoid treating embedded diff whitespace as repository errors

This PR description was generated by Pi using gpt-5.6-sol

@vercel

vercel Bot commented Aug 31, 2026

Copy link
Copy Markdown

The latest updates on your projects. Learn more about Vercel for GitHub.

1 Skipped Deployment
Project Deployment Actions Updated (UTC)
hunk-web Ignored Ignored Aug 31, 2026 12:45pm

Request Review

@greptile-apps

greptile-apps Bot commented Aug 31, 2026

Copy link
Copy Markdown
Contributor

Greptile Summary

The PR establishes a repository-level convention for preserving bounded engineering experiments and archives the discarded session-broker Effect lifecycle experiment without changing production behavior.

  • Adds the experiment plan, final report, and navigation documentation.
  • Retains control and treatment measurements, inventories, reconstruction patches, validation metadata, and checksums.
  • Configures Git and oxfmt to preserve generated evidence byte-for-byte and adds an empty Changeset.

Confidence Score: 5/5

The PR appears safe to merge because it only adds documentation and internally consistent archival evidence, with no production behavior changes.

The retained manifest verifies successfully, the Git attributes preserve patch bytes as intended, and no concrete build, runtime, security, or archive-integrity defect remains.

Important Files Changed

Filename Overview
research/README.md Defines a clear repository convention and required evidence boundary for archived experiments.
research/session-broker-effect-lifecycle/report.md Archives the experiment’s measurements, limitations, defects, and discard decision without changing runtime code.
research/session-broker-effect-lifecycle/artifacts/MANIFEST.sha256 Covers the retained archive, and every listed path and digest verifies successfully.
.gitattributes Marks artifact content as generated and preserves reconstruction patch bytes by disabling text normalization.
.oxfmtrc.json Excludes byte-exact generated evidence JSON from automatic formatting as described by the archive convention.

Reviews (1): Last reviewed commit: "docs(research): archive session lifecycl..." | Re-trigger Greptile

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.

1 participant