|
1 | 1 | # Licensing |
2 | 2 |
|
3 | | -This ecosystem separates three things: code, documentation, and data — and, within |
4 | | -data, records the project can openly license from third-party material it cannot. |
| 3 | +Open by default. Everything this project releases is openly licensed. |
5 | 4 |
|
6 | 5 | ## Code and tooling — Apache-2.0 |
7 | | -All software here (validation toolkit, evaluation harness, release tooling, schema) |
8 | | -is licensed under Apache-2.0. See `LICENSE`. |
| 6 | +The validation toolkit, evaluation harness, release tooling, and schema are licensed |
| 7 | +under Apache-2.0 (`LICENSE`). |
9 | 8 |
|
10 | | -Rationale: a permissive licence with an explicit patent grant, universally accepted, |
11 | | -that places no barrier on adoption by researchers, institutions, or industry. |
| 9 | +Rationale: a permissive licence with an explicit patent grant, universally accepted, that |
| 10 | +places no barrier on adoption by researchers, institutions, or industry. |
12 | 11 |
|
13 | | -## Documentation — CC BY 4.0 |
14 | | -Documentation and schema definitions are licensed CC BY 4.0. |
| 12 | +## Data — CC BY 4.0 |
| 13 | +All released sentence pairs and their metadata are licensed CC BY 4.0 (`LICENSE-DATA`). |
| 14 | + |
| 15 | +Rationale: the standard open-data licence for research corpora. It permits reuse, |
| 16 | +redistribution, and commercial use with attribution — maximising downstream adoption |
| 17 | +while ensuring the ecosystem and its contributors are credited. |
15 | 18 |
|
16 | | -## Released data — CC BY 4.0, for cleared records only |
17 | | -Sentence pairs are released under CC BY 4.0 **only where the project holds or has cleared |
18 | | -the rights to redistribute them** — records original to the project, in the public domain, |
19 | | -or drawn from CC-compatible sources. See `LICENSE-DATA`. |
| 19 | +## Documentation — CC BY 4.0 |
| 20 | +Documentation, dataset cards, and schema definitions are licensed CC BY 4.0. |
20 | 21 |
|
21 | | -Records from third-party sources under their own licences are redistributed under those |
22 | | -licences with attribution, or referenced by pointer where redistribution is not permitted. |
23 | | -**Every release manifest records the licence and attribution for each source it includes.** |
24 | | -No record is released under CC BY 4.0 unless its provenance permits it. |
| 22 | +## Attribution |
| 23 | +Cite the corpus using `CITATION.cff` and the release DOI. Attribution is required under |
| 24 | +CC BY 4.0 for both data and documentation. |
25 | 25 |
|
26 | 26 | ## Data governance |
27 | | -Data governance follows the CARE Principles for Indigenous Data Governance alongside FAIR. |
| 27 | +Governance follows the CARE Principles for Indigenous Data Governance alongside FAIR, |
| 28 | +ensuring the communities whose languages are represented share in the benefit and retain |
| 29 | +a voice in how the data is stewarded. |
| 30 | + |
| 31 | +## Scope boundary |
| 32 | +These open licences cover the public ecosystem only — released data, schemas, |
| 33 | +documentation, benchmarks, validation utilities, and contribution infrastructure. |
| 34 | +Proprietary NKENNEAi assets (model weights, commercial APIs, customer data, and |
| 35 | +confidential training methods) are separate and are not covered here. |
0 commit comments