Skip to content

Disputed boundaries policy

Any repository of national boundaries will eventually receive a pull request that is really about sovereignty. This page exists so that the answer is written down before the first one arrives, rather than improvised under pressure.

The policy

Boundaries follow the upstream source named on each dataset's page. This project takes no position on sovereignty.

Four consequences:

  1. The source is always disclosed. Every dataset page and every manifest.json names where the geometry came from and when. A boundary in this repository is a statement about that source, not about who is right.
  2. Geometry is not edited to reflect a claim. Not the maintainer's, not a contributor's, not a government's. If the upstream source draws a line somewhere, that is where the line is.
  3. Disagreements are documented, not resolved. Where a boundary is contested, the dataset's notes field records that fact and identifies the competing claims. Users are told there is a dispute; they are not told who wins.
  4. Alternative claims may be published side by side. Where a well-sourced alternative delineation exists, it can be added as a separate dataset with its own source attribution — never by overwriting the existing one.

Why not just "use the UN position"

It is a reasonable suggestion and it does not work. The UN itself does not publish authoritative boundary geometry for disputed areas; its cartographic products carry disclaimers precisely because member states disagree. Adopting "the UN position" would mean choosing an interpretation and calling it neutral.

Deferring to a named, dated, checkable source is a weaker claim, and that is the point. It is one this project can actually honour.

What this means for contributors

Acceptable:

  • "This boundary does not match the source cited in the manifest." — a factual claim about the data, checkable, and if correct, a bug.
  • "Source X has published a revision dated Y." — an update.
  • "This area is disputed and the manifest does not say so." — a documentation gap worth fixing.

Not acceptable:

  • Editing geometry to reflect a territorial claim.
  • Changing a source attribution to one that draws the preferred line, without a substantive reason for preferring that source.
  • Issues arguing the merits of a claim. These will be locked; see the Code of conduct.

Cases already in scope

Chile's existing data contains two:

  • The Chilean Antarctic Territory. Región de Magallanes y Antártica Chilena is present at ADM1. Chile's Antarctic claim overlaps those of Argentina and the United Kingdom, and the Antarctic Treaty suspends all such claims. The geometry reflects what BCN publishes.
  • Isla de Pascua / Rapa Nui. Not disputed between states, but the naming is contested locally between the Spanish and Rapa Nui forms. The shapeName field carries the source's form; alternative names can be recorded separately.

Neither is currently annotated in a manifest, because no manifests exist yet. Both will be when the data is restructured.

Naming

Place names are frequently as contested as the lines.

  • shapeName carries the name used by the upstream source, in the local language.
  • shapeNameEn may carry an English exonym where one is in common use.
  • Alternative or contested names belong in notes, with their context.

The project does not rename features to prefer one community's form over another's. Where a name is genuinely contested, record both and say so.

Changing this policy

Via an issue and public discussion. It is deliberately mechanical: a policy that requires judgement about who is right is a policy that will be relitigated on every pull request.