A month or so ago, I posted about the NISO document for building digital collections. Buried in that document was a great quote about the different kinds of metadata.
Essentially, NISO claims, it breaks down to three types:
- Descriptive Metadata is what we normally think of as content modeling. This is metadata that describes an object – the properties that define what it is: title, body, author, etc.
- Administrative Metadata is not something that the public consumes, but rather someone that is used to manage the object. publish date, expiration date, rights management, etc. I imagine that templating data would fit in here too – number of columns, color scheme, etc. – as would anything else that does something to the content that the user doesn’t consume directly.
- Structural Metadata is what I would call “relational content modeling." This is the information that relates this content to other content, and makes it possible for this content to fit into larger structures, and provides functionality not unlike your average relational databases.
Of course, all this depends on whether or not you think “metadata” is even a thing at all, and I’ve argued in the past that metadata in web content management is an outdated concept.