This is a set of best practices for writing and parsing metadata to handle attribution and license information.
The purpose is to provide an area for discussion and consensus-building about how to best manage metadata that describe how to credit the creators of digital works.
The technologies and standards available today provide many different mechanisms for doing this. There are both formal and ad hoc schemas for recording the metadata, and many different ways to make the information available as embedded or linked metadata. In addition, tool support for this metadata varies a lot, and many metadata sources misuse standards or just use partial standards.
These best practices are intended to handle these complexities by providing a set of rules covering:
- How to write metadata that is robust and can be understood by many applications and services.
- How to parse existing metadata to extract as much information as possible, even when it is ad-hoc or patchy
There's one file per main topic, all written with Github Markdown syntax.
Example files may later be added in a subdirectory.
All thoughts are welcome! You can contribute in several ways:
- Comment on the pages for the documents in the repository. It is also possible to comment on individual lines in the documents.
- Write an issue to suggest changes
- Fork this repository and change the documents yourself. Then submit a pull request to get it merged into the master.
A mailing list is planned to provide a forum for discussions about metadata best practices. This section will be updated when it is available.
For the purpose of these practices, credit metadata consists of any combination of this set of four parts:
- Work information
- Creator attribution
- License information
- Source works
The three first parts can have both a textual representation and a URL that links to more information.
The list of source works is ideally the credit metadata for the source work, and can thus have source works of its own. How many ancestor works to include in different scenarios is a topic of these best practices.
These best practices does not discuss how to encode license information itself, beyond a set of standard URIs (e.g. the Creative Commons license deeds). How to guide or even force users to comply with the license of the work when redistributing or remixing works is not covered here either.
Properties are written with the common namespace prefix,
e.g. dc:title.
Namespace definitions are generally excluded from inline examples (but not from standalone example files). The following set of namespaces are used, including some de facto namespaces that are only identified by their prefixes:
| dc: | Dublin Core elements | http://purl.org/dc/elements/1.1/ |
| dcterms: | Dublin Core terms | http://purl.org/dc/terms/ |
| cc: | ccRel | http://creativecommons.org/ns# |
| xhv: | XHTML | http://www.w3.org/1999/xhtml/vocab# |
| og: | Open Graph | http://ogp.me/ns# |
| twitter: | Twitter cards | N/A |
| flickr_photos: | Flickr metadata | N/A |
The repository was initiated by Commons Machinery: http://commonsmachinery.se/
libcredit: A multi-language implementation of these best practices. https://github.com/commonsmachinery/libcredit