Skip to content

Ecosystem Direction Decision Gate

Status: G81 decision artifact. No G82 integration track is open yet. This page compares possible ecosystem directions against the evidence from G75-G80 and keeps implementation blocked until adoption signals justify one track.

Entry evidence available now

Evidence Status
Release/adoption readiness Local v0.2.0-alpha.0 candidate docs, package metadata, and release-prep dry-run are ready; no tag has been published by a human owner.
Dogfood pilot Synthetic YAML pilot plus C++ robot-cell app exercise runtime semantics without external dependencies.
Compatibility harness Stable-v0.2 header/API, schema, CLI JSON, metrics, trace, live observe, and package smokes exist.
Packaging hardening Runtime-only/default/CPack/vcpkg/Conan draft package matrix exists; registry publication remains deferred.
Reliability/adoption loop Bounded reliability tiers and feedback/triage docs exist; no real external adoption data is present yet.

Candidate G82 tracks

Track Candidate Upside Main blocker Current decision
G82a Production observability exporter Converts metrics/trace/live observe into real telemetry integration. Requires exporter dependency/transport/API decision and evidence that users need it. Deferred.
G82b Native Python binding Improves scripting and notebooks. Requires ownership/performance/API stability decision; current CLI-backed preview may be enough. Deferred.
G82c Real ROS 2 adapter Attractive robotics adoption path. Requires ROS dependency/version/executor/lifecycle boundary and hardware-like validation plan. Deferred.
G82d Editor/LSP Better graph-authoring UX. Requires proof JSON Schema + diagnostics are insufficient and a maintenance plan. Deferred.
G82e Package registry publication Improves install/adoption with least runtime-scope risk. Requires human release owner, exact release artifacts, credentials, and clean-machine evidence. Recommended future first track after a human v0.2.0-alpha.0 release decision.

Recommendation

Do not implement any G82 track in this goal. If adoption evidence appears after G75-G80, prefer G82e package registry publication first because it improves access without adding runtime dependencies or widening API semantics. Keep G82a-d behind explicit user demand or concrete downstream integration issues.

Decision criteria for opening a track

Open at most one track when all of these are true:

  1. a human owner accepts the scope and credentials/dependencies involved;
  2. at least one adoption signal names the need, such as install friction, telemetry integration need, ROS integration request, Python automation limit, or editor authoring friction;
  3. the track has scope, allowed files, non-goals, acceptance, validation, and blocker handling before implementation;
  4. the track preserves G73 live observe hot-path invariants and G74 generated showcase anti-rot rules;
  5. status/backlog/changelog/release docs make preview/deferred claims explicit.

Safe independent work before any track

  • keep examples, package, reliability, and adoption gates green;
  • collect real issue/PR feedback through the G80 triage workflow;
  • improve docs and package smoke evidence without publishing;
  • write blocker notes for any dependency/API decision that would otherwise be guessed.