License and redistribution terms
CantHarm separates open release artifacts from materials whose redistribution may depend on source-specific permission.
Field-level policy
| Material | Policy | Public release note |
|---|---|---|
| Workbook release artifact | CC BY 4.0, subject to the dictionary-derived source-text caveat. | Public release license wording. |
| Derived labels, source references, and metadata | Open release metadata. | Keep distinct from dictionary-derived definitions/source text. |
| Dictionary-derived definitions or source text | Outside open-license scope unless source-specific permissions are confirmed. | Do not describe all source text as unconditionally CC BY. |
| Audio clips | CC BY 4.0 with speaker consent. | Accompanied by responsible-use guidance against voice misuse. |
| Code | MIT License. | Separate from dataset artifact licensing. |
Dictionary-derived source-text caveat
Labels, metadata, and source references can be documented as release metadata. Dictionary-derived definitions or source text should not be redistributed as open-license text unless the relevant source permission is confirmed.
Audio license note
The current audio release is described as CC BY 4.0 with speaker consent. The site also documents unacceptable voice misuse, but that documentation is kept separate from the license grant to avoid contradictory license wording.
Acceptable-use note
Acceptable-use and ethics notes are provided as public release documentation. The public contact endpoint is yueyu_dimsum@163.com, with backup contact qijiayin@139.com.