Skip to main content

INVESTIGATE: NLOD attribution across everything Atlas publishes

Status​

Open — 2026-09-11. Split out of INVESTIGATE-all-brreg-organisations at Terje's direction: "the NLOD marking goes for all data in atlas. write that as a separate thing that we need to handle. so that we dont focus on that now."

Deliberately not being worked on. This file exists so the obligation is recorded rather than carried as a footnote inside a Brreg investigation, where it would have looked Brreg-specific and would have been closed with it.

The obligation​

Atlas republishes Norwegian public-sector open data. Every upstream it uses is NLOD-licensed, and NLOD requires attribution. That applies to all of it — SSB, FHI, Bufdir, Brønnøysundregistrene — not to any one source.

What already exists​

The machinery is built and mostly used:

  • every module under ingest/src/sources/ declares publisher, license, license_url and attribution in its manifest.yml
  • api_v1.meta_sources publishes publisher, license, license_url, attribution per source
  • so a consumer can already retrieve the attribution for any source they query

The gaps, as far as they are known​

  1. 🔴 seed-sources/brreg-enheter/ has no manifest. Only README.md and index.ts — no publisher, no license, no attribution. Atlas serves Brreg-derived data today with nothing recorded. This is live at 122 rows and does not depend on any decision about scale.
  2. Other seed sources are unexamined. seed-sources/ holds nine modules; only brreg-enheter has been looked at. The rest may have the same gap.
  3. Derived data is unresolved. marts.* and api_v1.* views combine several upstreams. Nothing currently says which attributions a given view inherits, and a consumer reading one endpoint has no way to know it is touching four licensors.

Open questions — why this is an INVESTIGATE and not a PLAN​

  • Where must attribution appear? Per-source in meta_sources only, or also in the OpenAPI info block, the docs site, and HTTP responses?
  • How does it compose? If api_v1.indicator_summary derives from SSB and FHI, is listing both sufficient, and where?
  • Does the frontend carry it? atlas-frontend consumes the public API as a reference implementation — if attribution is an obligation on republication, the reference implementation should demonstrate discharging it.

First step whenever this is picked up​

Audit, before designing: for every source in meta_sources, does it declare a licence and an attribution, and does any published surface show them. The answer for brreg-enheter is already known to be no, which is one concrete thing to fix regardless of how the rest is answered.