It is important to understand that metadata is often layered in complex publishing systems and in the content use-cycle. The diagram below illustrates at a high level, some of the tiers of metadata enrichment applied to published content.

Blocks in green at the top of the hierarchy are owned and enabled by systems typically outside of the publisher's control - certainly outside of editorial and eBook distribution. They include enrichment of the content by systems being operated by end users - both consumers (including non-paid experts) and experts, who are paid to enhance content for easy display, retrieval, sale, etc. As examples:
- A non-professional consumer inherently enriches published content when he or she looks at it (tracked by the system), annotates it, writes a review, forwards it to a friend, etc.
- A paid expert enriches published content when he or she adds an abstract, additional expert-tagging, etc to enable users to find and purchase relevant content.
It is important to note that this enrichment is stored as a separate piece of XML data –acting as metadata to enrich the original content.
Blocks in blue starting from the bottom of the hierarchy are owned outright by the publisher. It includes XMLs specifically for use in their editorial process and standard XMLs, like DocBook for publisher system inter-operability, re-use, etc. Packaging XMLs, like ePub, are contained in this sphere as publishers generally create their own eBooks, often using third parties.
The third block in the hierarchy, Package Enhancements is gradated between blue and green because its current ownership is unclear. It involves adding enhancements embedded in the Packaging XML – adding hyperlinks to other content, video, etc. Publishers are actively doing this to top titles, and there is nothing that prohibits a third party with non-DRM access to a packaged XML from embedding such enrichment specific to their marketplace. The key difference between these enhancements and those in green blocks would be the change to the actual content container – which would make those changes automatically available to any user with access to that edition – as opposed to layered-on-top-of enhancements which are controlled by the annotation sharing system.
