XMP

When XARK is used to store information about a digital artifact (transcriptions, people, dates, etc.), it is preferable to store this information inside the artifact file itself. This helps to ensure that the evidence remains with the source as it is shared.

The XMP Standard (opens new window) (ISO 16684-1) was developed primarily by Adobe, and provides a means of storing XML+RDF data within JPEG, PDF, WebP, and other common file formats. It is comprised of RDF statements encoded in XML. Since it relies on RDF's data model, XMP has a very "flat" hierarchy (unlike JSON). Each XMP "packet" can only contain a set of properties, where each property has a property identifier (a URI) and a value (several possible types). To accommodate this format, XARK data should be stored with separate XMP packets for each Artifact, Subject, Claim, and Note.

There must be at least one Artifact packet per file. There will usually be multiple Subject packets per artifact, and multiple Claim packets per subject. Note packets are optional and not necessarily expected.

Each property identifier can only appear once per packet. Multiple values for a given property are supported using unordered arrays, ordered arrays, or structured values.

Because we are not storing the entities in a hierarchy, there are two additional required properties that would not be required in hierarchal formats such as JSON (where these are known from the hierarchy context):

TIP

All property names in XMP are in kebab-case (lowercase, dash-delimited). This differs from JSON, where they are rendered in pascalCase.

Required Properties

Class

To identify whether the XMP Packet represents an Artifact, Subject, Claim, or Note. Valid values:

• https://xark.org/xark#artifact
• https://xark.org/xark#subject
• https://xark.org/xark#claim
• https://xark.org/xark#note

Parent

To identify the XARK ID of the parent of this entity. Each Note links back to its parent Claim, each Claim to its parent Subject, and each Subject back to its Artifact.

::: warn Values are not permitted to be arrays. While RDF allows for this, it is not compatible with the XARK itself. If there are multiple "truths" for a given attribute (for example, someone with multiple middle names), these should be stored as separate Claims. :::