A key design objective for user annotation is to allow analysts, data providers, and even external collaborators to decorate metadata objects in the Data Catalog and have those decorations propagate to records that are linked downstream of the annotated item. For instance, adding a comment to sample should mean that users working with that samples measurements and files, or even derived data products, will have access to that sample-level annotation.
There are currently two forms of user-initiated annotations currently implemented. Tags are short alphanumeric strings drawn from a controlled vocabulary, while Text Annotations are free text documents with a an optional subject line. Support for an Attribute-Value-Unit class of annotation is forthcoming.