Skip to content

BACKLOG.md

Latest commit

 

History

History
124 lines (81 loc) · 4.62 KB

BACKLOG.md

File metadata and controls

124 lines (81 loc) · 4.62 KB

BACKLOG.md – Rust-as-Spec Platform Cell

Purpose: Ideas that are interesting but premature for the kernel. These features are held here until forks demonstrate real need and validate the design.

Policy: Implement backlog items ONLY when:

  1. At least one fork has a concrete use case
  2. The feature is generically designed (works for any service)
  3. An ADR is written justifying the addition to the kernel
  4. It's implemented with must_have_ac: false initially
  5. Once proven in 2+ forks, consider promoting to kernel

Review: Revisit quarterly based on fork feedback and emerging patterns.

Bundle System Enhancements

New Bundle Types

  • troubleshoot_ac – Focus on a failing AC, include related tests/specs/docs
  • implement_task – Gather task dependencies, related ACs, and implementation hints
  • explain_decision – Reconstruct reasoning behind an AC or design choice from ADRs and commits
  • measure_coverage – Show test/doc gaps for a given AC or REQ

Bundle Quality Features

  • Bundle caching and incremental updates (only regenerate changed files)
  • Priority-based file inclusion (exclude low-signal artifacts if bundle exceeds size limits)
  • AC coverage signals in bundles (explicit "tests this AC", "documents this AC" flags)

Graph and Visualization

Advanced Graph Rendering

  • Mermaid graph with filtering (by REQ, AC status, owner, label)
  • Interactive graph explorer in /ui (click nodes to expand/collapse)
  • Dependency visualization for tasks (task → related ACs → tests → docs)
  • Impact analysis (which tests/docs/ACs would fail if AC X changes)

Release Evidence and Notes

Automated Release Artifacts

  • Changelog generation from evidence bundles (group changes by category)
  • Automatic Keep a Changelog formatting (ADDED, CHANGED, DEPRECATED, REMOVED, FIXED, SECURITY)
  • Release delta reports (diffs between version bundles, what changed and why)
  • Evidence-based release notes (tie each change to REQ/AC/test coverage)

Agent Hints and Suggestions

Intelligence for Task Selection

  • Effort estimates for tasks (based on AC complexity, test count, breadth of changes)
  • Dependency-aware prioritization (suggest blocking tasks first)
  • Context-aware bundle recommendations (suggest bundle type based on task type)
  • "Next best AC to implement" algorithms (considering skill gaps, dependency chains, impact)

DevEx Tooling

Interactive Developer Experience

  • cargo xtask repl – REPL for exploring spec_ledger, running queries (e.g., "show ACs for REQ-X")
  • cargo xtask watch – Auto-run tests when spec changes, show live feedback
  • IDE integrations (LSP for spec_ledger.yaml with completions, linting, validation)
  • VS Code extension for AC navigation (inline previews, test linking, bundle shortcuts)

Testing and Validation

Advanced Test Insights

  • Mutation testing for AC coverage (ensure tests would fail if AC code mutates)
  • Property-based testing scaffolds (generate test data from AC assertions)
  • Test gap analysis (identify ACs with low test diversity or missing edge cases)
  • Flakiness detection and reporting (flag tests that pass/fail unpredictably)

Documentation

Automated and Intelligent Docs

  • Auto-generated API docs from spec_ledger (keep docs fresh with spec changes)
  • ADR dependency graphs (show how design decisions relate and evolve)
  • "Why does this exist?" explanations for each AC (rationale, history, alternatives)
  • Cross-repo template documentation portal (shared learnings across forks)

How to Move Items from Backlog to Kernel

  1. Fork demonstrates need – A downstream project (Knowledge Hub, etc.) has concrete use case
  2. Generic design – Feature works for any Rust-as-Spec service, not domain-specific
  3. ADR written – Architecture decision recorded with rationale and tradeoffs
  4. Pilot implementation – Code merged with must_have_ac: false (optional in selftest)
  5. Validation – Feature proven useful in 2+ independent forks
  6. Promotion – Move to kernel as must_have_ac: true, document in ROADMAP.md

Explicitly Out of Scope

These are not candidates for the kernel backlog:

  • Domain-specific features – Calculators, content management, business logic
  • Service integrations – Payment providers, auth systems, vendor SDKs
  • Application-level concerns – UI frameworks, deployment orchestration
  • Anything that couples the kernel to a specific business domain

Forks should implement domain features themselves; the kernel stays focused on governance, specs, and workflows.

Last Updated: 2025-11-26 Quarterly Review Scheduled: 2026-02-26