OntoBoard
Run your whole ODK and ROBOT pipeline inside Protégé. No terminal, no make, no Docker -
for the projects OntoBoard makes. A repository generated by ODK itself builds through Docker,
Podman or a native ODK install; what needs what.
A Protégé Desktop plugin for people who maintain OBO-style ontologies: scaffold an ODK project,
build it, run ROBOT's report, explain what the reasoner concluded, refresh your imports and cut a
release — all from the editor you already have open, against the ontology already in front of you.
And a board that draws the whole ontology, the way a whiteboard would.
New to ontologies, ODK or both? OntoBoard → Start here: the guide... walks you from an idea to a
released ontology, one step at a time — or read the guide first.
Why
The ODK is excellent and it lives in a container. The usual day looks like this: edit in Protégé,
switch to a terminal, sh run.sh make test, read a log, switch back, find the term, fix it, run it
again. Every loop crosses a tool boundary, and on Windows it crosses Docker as well.
OntoBoard embeds ROBOT itself — not a wrapper around a binary, the library — so the same
operations run in-process, on the ontology Protégé has open, and report into a dialog you can read
and save. A project OntoBoard scaffolded runs seven of its eight generated build targets with
nothing installed at all: no make, no robot, no shell, no container, on any platform.
Where a container genuinely is needed, OntoBoard uses the one your project declares — or Podman,
or a native odk install environment. OntoBoard → Check requirements tells you what
you can do right now and what to install for the rest.
Install
- Download
ontoboard-<version>.jar from the latest release.
- Delete any older
ontoboard-*.jar from Protégé's plugins/ folder. Two copies will not
resolve.
- Drop the jar into
plugins/ and restart Protégé.
- Open the board with Window → Tabs → OntoBoard: every term of the open ontology is on it.
Protégé Desktop 5.6.x or 5.5.0, both smoke-tested in the real host on every release. Java 8+
(Protégé bundles its own). The jar is large because ROBOT and its dependencies are inside it.
What it does
Project — scaffold a new ODK project from a wizard - empty, or from the ontology you have
open - or open an existing one. Turn an ontology in any format, with its earlier versions, into an
ODK repository whose git history has a release and a tag for each version, ready to push to
GitHub. Check what tooling you have. Edit ID ranges. Audit and refresh
imports. Clone a project from GitHub. Turn a competency question into a test by picking the terms.
Terms from other ontologies — find terms by label in an ontology given by its web address, a
file, or the page about it, and add them as an import. Imported terms shows which term came from
where, with its label, and adds or takes out terms, rebuilding the import at once.
A knowledge graph from spreadsheets — one ROBOT template per kind of thing, identifiers
generated, names checked across sheets, and each sheet turned into individuals in a graph that
imports your ontology. How
Build and release — run the project's build targets, in-process where possible, in its own
container or a native ODK environment where not. Compare two releases axiom by axiom. Stamp
provenance. Obsolete a term properly.
ROBOT, embedded — report, measure, profile, explain, SPARQL, transform, export terms, template,
rename IRIs, extract import modules. Nine of these are driven end to end in the real Protégé host
on every release, and nine have been compared against the real robot command to make sure the
answers match.
A board of the whole ontology — every class, property and individual as soon as the tab opens,
laid out as an outline: design patterns framed along the top - imported, or merged in and
recognised by their terms - classes as a tree, properties in
sections of their own. Drawn from the axioms you actually wrote — restrictions, unions, domains and
ranges, disjointness, property chains, SWRL rules: every logical axiom of two published BFO-based
ontologies. Select a term and its relations come forward while the rest fades; fold a class; jump
between frames; press Tab for a new subclass. Switch on the reasoner's conclusions, then right-click
a dotted edge and ask why — it answers with the axioms that force it.
History and reuse — every release of the project, what changed in each term by term, any of
them opened in a window of its own. Find terms in another ontology by their labels and import
them without looking up an IRI.
162 design patterns, built in — 122 from the ODP portal plus 40 extracted from MWO, PMDco
and NFDI MatWerk, each as OWL and Turtle. Browse by collection or publisher, or let the plugin
suggest the ones whose vocabulary your open ontology already speaks — with the reason, not just
a ranking — and import the terms you want through ROBOT's module extraction.
Notes and provenance — editorial notes that travel with the ontology, discussion links to your
tracker, and authorship stamped on what you change.
Collaboration — live shared sessions with cursors, or git: status, stage, commit, pull, push,
branch, without leaving the editor.
Documentation
ebrahimnorouzi.github.io/ontoboard — the full
documentation site. The same pages are readable here in docs/:
Status
Early and actively developed, but held to a standard: 2,660 unit tests, every release
smoke-tested in both Protégé versions, Windows checked in CI, the plugin self-checks on startup,
and each release ships a receipt in
protege-plugin/tools/smoke-receipt/ recording what was
verified and what was not.
Developed by Ebrahim Norouzi at
ISE / FIZ Karlsruhe.
Licensed under Apache 2.0.
