Author

Miko
Parkkinen.

Creator & maintainer of Metriplane

Metriplane is open-source robotics incident replay and regression-testing software for bounded workcells. It turns recorded runs into readable incident reports, verified evidence, and repeatable checks.

Project context

One project.
Two distinct records.

Metriplane v0.4.1 is the current Python distribution and CLI package. Metriplane v0.2.0 is the DOI-archived public release prepared as the primary software artifact for a manuscript submitted to SoftwareX. The site keeps those roles separate rather than treating the latest package as retroactive evidence.

Maintainer responsibilities

From the first recording
to a reproducible release.

Research-software release discipline

DOI release, citation metadata, reproducibility guide, claim evidence table, and bounded limitations.

Python packaging and delivery

PEP 517 metadata, PyPI distribution, version synchronization, optional dependency groups, and documented wheel boundaries.

Trusted release automation

OIDC publishing, TestPyPI verification, separate owner-started production promotion, strict artifact validation, and production install verification.

Recorded-run analysis

Planar state, process rules (domain packs), event timelines, incidents, Incident Reports, and evidence bundles.

Verification and testing

Archived author-run maintainer-gate evidence, deterministic replay checks, bundle verification, and repeatable regression checks.

Historical and repository-only integrations

ROS 2, Omniverse, dashboard, exporter, replay, and operator-review work is documented with its narrower scope; it is not part of the v0.3.0 demo-support claim.

Documentation architecture

Reviewer quickstarts, paper evidence docs, runbooks, schema notes, and public claim boundaries.

Measurement claim discipline

The preprint separates physical mapping, latency, continuity, replay, overload, and backend claims rather than substituting one for another.

Scope and claim discipline

The release separates current package behavior, historical research evidence, unsupported environments, and explicit non-claims.

Links

Next step

Explore the work at its source.

Read the implementation and release history, or send a specific technical observation.