Profile stability matrix¶
Canonical source for --profile / profile= behavior. Other pages link here — do not duplicate stability labels inline.
See Install and channels and Release status.
Published v1.0.0¶
| Profile | Status | Engine | Production recommendation |
|---|---|---|---|
rdfs |
Stable | ontologos-rl (rdfs module) → reasonable |
Embed RDFS materialization |
rl |
Stable | ontologos-rl → reasonable |
OWL RL saturation |
el |
Stable | ontologos-el (in-house) |
OWL EL taxonomy |
auto |
Stable | Detect → EL, RL, or DL | Prefer explicit profile when known |
dl |
Stable | ontologos-dl |
OWL 2 DL — gated HermiT catalog @ 30s |
swrl |
Stable | ontologos-swrl |
DLSafe SWRL subset |
dl-preview |
Preview | ontologos-dl (gated) |
No — use dl |
alc |
Preview | ontologos-alc |
No |
ql (detection only) |
Detection only | None | Use ELK or another QL reasoner |
What “stable” means here¶
- Stable: Suitable for production embedding within OntoLogos’s mapped construct subset.
dlpasses the in-scope HermiT gate; validate on your corpus — see Evaluator scope. - Preview: Explicit gating, incomplete rules, or
PreviewLimit/NotImplementedon common paths.
What “100% HermiT parity” means for dl¶
parity_pct = 100% counts only 889 in-scope catalog cases (461 Java + 428 WG), not all 1019 HermiT-derived entries. 130 Java cases are documented out of scope (internal, excluded, migrated). Tier C taxonomy checks allow OntoLogos to be a sound superset of HermiT, not identical output.
For the full breakdown, see Evaluator scope.
CLI quick reference¶
ontologos materialize ontology.owl # explicit RDFS (stable)
ontologos classify --profile rl family.owl # stable
ontologos classify --profile el pizza.owl # stable (after download.sh for Pizza)
ontologos classify --profile auto ontology.owl # stable routing
ontologos classify --profile dl ontology.owl # stable — gated catalog parity
ontologos classify --profile swrl ontology.owl # stable (DLSafe subset)
Related¶
- Evaluator scope — what 100% does and does not mean
- Preview profiles — ALC and
dl-previewlimitations - Comparison — vs ELK, HermiT, reasonable
- Choosing an API — Rust entry points
- Evaluator playbook — 30-minute evaluation