Replies: 1 comment
|
Thanks for the suggestion - we will add it to the list of discussion topics for the TSC. |
0 replies
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
RDF is the usual language for specifying subject-predicate-object statements about data. It's widely embedded in a variety of types of documents through XMP, which is a profile of it, to make metadata statements about the document (title, author, license, etc.). XMP is designed to be parsed out of a document without knowing the file format due to using certain fixed headers (XML processing instructions).
Supporting RDF as a syntax for this data would allow it to be embedded in a variety of documents, including PDFs, images, and XML (including XHTML) rather than having to embed a custom schema. It would be even better if it worked with the existing Creative Commons License Metadata schema, which is already commonly used for making license statements.
Right now, everyone has to parse RSL content in addition to other XML content, and it doesn't embed in non-XML documents. With RDF, RSL metadata can be embedded in any document and automatically parsed out, without having to care about the data. Could you consider adding an RDF schema for this information?
All reactions