Licensing
Three licences, because three different things live here and they have genuinely different reuse profiles. Conflating them is the usual mistake.
| What | Where | Licence | |
|---|---|---|---|
| Code | web/, tooling |
Apache-2.0 | LICENSE |
| Directory data | data/ |
CC0 1.0 | data/LICENSE |
| Documentation | docs/ |
CC BY 4.0 | docs/LICENSE |
Why each one
Code — Apache-2.0. A patent grant, and an explicit non-grant of trademark. The second
matters more than usual here. "NavCom" carries a promise about behaviour — that Distress
ends in a human or reports that it couldn't, that the directory shows its age rather than
guessing. Apache section 6 means you can fork the code and cannot take the name with it. A
careless fork trading on that name is a real hazard when the name is the safety claim.
Directory data — CC0 1.0. The directory is a public good or it isn't. A city, a 211
service, a mutual aid group or a street medic should be able to use corrected data with
zero friction and no lawyer. Anything less permissive — including NonCommercial, which
sounds protective — would block exactly the reuse
docs/product/propagation.md names as a growth mechanism.
CC0 is also the honest choice about what is being claimed. Facts about a shelter's opening hours may not be copyrightable at all in many jurisdictions, and asserting a right that may not exist would be its own small overclaim.
Contribution credit survives this. CC0 does not stop NavCom crediting the callsign in
verified_by, and docs/product/identity.md makes contribution
one of the two axes of standing. What CC0 gives up is legally compelling downstream
reusers to carry that credit. That is a price worth paying for a shelter list that anyone
can use.
Documentation — CC BY 4.0. These are argued documents rather than reference tables, they get quoted, and attribution is reasonable to ask for.
What no licence does
Every licence here disclaims warranty, so that much is covered. None of them tells a human that the data is time-sensitive, and that is the failure this project is most exposed to, because someone acts on it at 10pm outside a closed door.
So, in plain language and not in legal language:
This directory is maintained by volunteers. It shows how old every perishable fact is, on purpose. When a check is too old to trust, it says "call first" — that is a real answer, not a missing one. A blank means nobody has established something, never that there is no restriction. Please call ahead before sending anyone anywhere.
There is no organisation behind this
NavCom is not a business, not a non-profit, and not a registered entity. Nobody is trying to make money from it and there is no company to sue, sponsor, or join.
That is not a disclaimer — it changes what these licences actually do, in three ways worth stating plainly.
Copyright rests with contributors, not an organisation. There is no entity to hold it and none is needed. Each person owns what they wrote and licenses it to everyone else under the terms here. No CLA, no assignment, no paperwork.
The trademark clause is intent, not enforcement. Apache section 6 withholds trademark rights, but with no entity there is no registered mark. Common-law rights from use are weak and jurisdiction-bound, and enforcing them takes standing and money that do not exist here. The request stands and rests on good faith: fork the code, and call it something else, because the name is a claim about behaviour.
It is why the data is CC0 rather than share-alike. A copyleft data licence only protects anything if somebody will enforce it. Choosing one here would impose real friction on a 211 service or a city — the exact reuse this project wants — in exchange for a protection that exists only on paper. Permissive is the honest choice when enforcement is not available.
What this means for data from elsewhere
CC0 has a consequence worth naming, because it constrains how the directory can grow.
OpenStreetMap is ODbL, which is share-alike for databases. Extracting names and addresses from it in bulk would make this a derivative database and pull the whole directory out of CC0. So OSM is used as a candidate generator — it proposes places that might matter, a person establishes the facts, and only what was independently established is published. Candidates are a working artifact and never become published data.
Records enter resources.csv sourced from the organisation's own site, government open
data, or a person who went there. Government open data is public domain and strong in the
United States, patchy elsewhere; OSM is global and ODbL. No single source is both, and
pretending otherwise is how a licence gets quietly broken.
Contributing
By contributing you agree your contribution is licensed under whichever of the three covers the files you touched. You keep your own copyright — there is no entity to assign it to and none is wanted. No CLA, no paperwork.
Never contribute anything about the people being served — no licence makes that
acceptable, and it is refused rather than edited. See
CONTRIBUTING.md.