Features & roadmap

Bioregistry, OBO Foundry, OLS, TIB, LOV and NeuroSchema, merged into one record per resource, with every field traced to its source.

prefix.cc, w3id.org, Zenodo, GitHub and suggested URLs are probed by content negotiation, and only real ontologies, vocabularies and schemas are accepted.

Each registry download is archived with its checksum, so the catalog can be rebuilt offline, and a single registry can be refreshed on its own.

The submission form opens an issue that a bot turns into a pull request, with link, DOI and registry checks on every change.

A check whose only evidence source doesn't cover a resource is "not assessed" instead of failed, and no problem is penalised twice.

ROBOT (report, ELK reasoner, OWL 2 DL profile) and OOPS! pitfall scans run on every ontology, with the OBO Dashboard shown side by side for comparison.

Reference DOIs are checked against Crossref, DataCite, OpenAlex, Semantic Scholar and doi.org.

ROBOT (report, ELK reasoner, OWL 2 DL profile), pitfall scans and import reading run on every ontology with a download, several at a time. A full pass resumes where it stopped, and each run reports how many ontologies have been checked.

owl:imports and the terms an ontology uses are read from the file itself, so upper-ontology alignment (BFO, COB, DOLCE, GFO, gUFO), RO use and vocabulary reuse are measured for every checked ontology, not only those listed in OLS or Bioregistry (rubric v4.1).

A published report of how sensitive grades are to the weights (random and one-at-a-time changes), how well our quality checks agree with the OBO Dashboard, and how scores compare with expert ratings.

Each resource's page shows how its score, every criterion and the quality metrics changed from one evaluation run to the next.