# Licence of the data

This applies to the published dataset — `programme.parquet` and everything derived
from it. The code in this repository is licensed separately, under the MIT licence in
`LICENSE_CODE`.

## The source

The data comes from the [Förderdatenbank](https://www.foerderdatenbank.de) of the
German federal government. Its
[imprint](https://www.foerderdatenbank.de/FDB/DE/Meta/Impressum/impressum.html)
licenses the texts of the site under
**[CC BY-ND 4.0](https://creativecommons.org/licenses/by-nd/4.0/deed.de)** and names
the **Bundesministerium für Wirtschaft und Energie** as the rights holder.

That licence passes through: this dataset carries no licence of its own to grant,
because it holds the ministry's texts. CorrelAid e.V. claims no additional rights
over them.

## What was done to the data, and why ND permits it

CC BY-ND forbids distributing modified versions. What this project does is
reformatting, not modification of the licensed texts:

- the XML export is transposed into one row per Förderprogramm;
- category values are reduced from CMS classifier paths to their leaf codes;
- the export's internal document references are resolved into contact fields and
  link lists;
- five category columns are *linked* to published code lists as SKOS matches.

No text is rewritten, summarised, translated or edited. Category codes are published
as the export states them and are never substituted for a code list's own — which is
the other reason for linking rather than replacing. Where a value could not be mapped
honestly, it is left unmapped rather than approximated.

The columns this project adds (`id_url`, `id_hash`, `on_website_from`,
`last_updated`, `previous_update_dates`, `deleted`) are observations about the
records, not content taken from the source.

## Attributing a single row

Every row carries a `license_info` column with the attribution to reproduce, for
example:

    Einstiegsgeld von Bundesministerium für Wirtschaft und Energie, lizensiert unter
    CC BY-ND 4.0 (https://creativecommons.org/licenses/by-nd/4.0/deed.de),
    Quelle: https://www.foerderdatenbank.de/FDB/Content/DE/Foerderprogramm/Bund/BMAS/einstiegsgeld.html

That column exists because dataset-level metadata is no help to someone holding one
extracted record. The same information is in the DCAT metadata at dataset level, as
`dct:license` and `dcatde:licenseAttributionByText`.

## No warranty

The dataset is provided as-is. It is a machine-readable rendering of a third-party
source, published in the hope it is useful for analysing the German funding
landscape; it is not an authoritative statement of any funding programme's terms.
Always check the linked source page before relying on a value.