search

c2pa-org / public-draft

Repository for the public drafts of the C2PA Specifications
Creative Commons Attribution 4.0 International
35 stars 1 forks source link

Provide some guidance on using DIDs as identifiers for assets #21

Closed lrosenthol closed 2 years ago

lrosenthol commented 2 years ago

From the W3C's Credential CG

But DIDs are expressly designed to identify anything -- including the 'asset'. Why limit it to just an 'actor'. DIDs were not designed merely to apply to 'actors' like people and machines. They also refer to specific assets like your 'image, video, document'. But you're not taking advantage of this, apparently.

As I was getting at above: how will you handle, say, a small publisher or self-publisher of assets which they are identifying using a system of DIDs. What if the publisher, or author, and each of the assets they produce, are all identified as a linked system of DIDs, so that a user refers back to the DID Document from any asset. Can this somehow be integrated into the C2PA system?

Absolutely! The Schema.org grammar for a CreativeWork – which is a component of the C2PA manifest – has a pre-defined field called (oddly enough) identifier which would be the logical spot for the asset’s DID to be recorded. You can use additional CreativeWork fields to refer to other referenced/linked works if you wished.

In addition, the DID could serve as the instanceID for a claim or ingredient.

hackerfactor commented 2 years ago

Section 7.1 uses "DID" as a unique document identifier. This seems to redefine the definition as used by XMP.

In XMP, DID is the document identifier (identifies the parent source). IID is the instance identifier (identifies the current save/alteration). With XMP, "save as" creates a new DID and IID, while "save" only changes the IID.

lrosenthol commented 2 years ago

The change here would be in 6.2, Other Identifiers. TWG agrees to add this.

lrosenthol commented 2 years ago

Such guidance is given in the second public draft of the document.