Repositories should be able to actively and in real-time collect and expose activity (e.g. information about changes, additions, comments, annotations, peer-reviews, accessess, downloads, etc.) pertaining to scholarly objects they host. Authors of the scholarly object involved in an activity, other repositories, and a variety of consuming applications that keep the pulse on scholarship as it happens should be able to receive metadata about activity not only retrospectively through harvesting, but also in real-time. To that end, notification mechanisms need to be put in place. Depending on the use case, these could be point-to-point notifications (e.g. an author is directly notified about a citation to her paper) or publish/subscribe notifications (e.g. a consuming application interested in peer-review subscribes to a channel on which review events are posted). In addition, value added services should be able to consume such activity information producing new notifications in turn. For example, this could be exemplified by academic recommender systems, which can, based on past (even anonymous) activity information, significantly help users in navigating research objects stored across repositories globally. In order to achieve such functionality, unique identification (by means of HTTP(S) URIs) of scholarly objects and actors (e.g. authors, reviewers, institutions) in the scholarly communication environment is essential.
In addition to the recommended technologies and standards, other messaging protocols (e.g. AMQP, Kafka) provide a common mechanism for communication between publishers of any kind of Web content and their subscribers. This behaviour will also require the development of standard vocabularies in order to expose user interaction data in a common format.