In contrast to single-record annotations, bulk annotations cover multiple, arbitrarily selected items and multiple, arbitrarily selected attributes of a dataset. The data to be annotated can be selected via a one-by-one selection of relevant items and fields or via the complete result set of a user's search query. Thus, a single annotation event can cover an enormous number of annotated fields. When multiple users or applications annotate data, and when annotations may have multiple dimensions it is no longer feasible to store all annotations per individual field.
To demonstrate the power of bulk annotations our web-based application MyDBLP extends DBLP by user-generated quality annotations, such as scores, comments or other dimensions. The prototype implements multiple storage schemes of bulk
annotations on arbitrarily selected sets of data.