In the context of the Web, a feed is a message that contains a representation of a sequence of references to resources that have been appended to or modified within their containing site.

This sequence contains at least the URI of the resource, its title, and often an excerpt or abstract of its content, if not the full text. Other properties of meta-data associated with these resources is often included, such as the publication date, a unique identifier, as well as other extensions, both standard and proprietary. Feeds often also contain a header that provides generic information about the source.

Open question: Is a feed a first-order resource, or is it derived from another resources? For example, a feed of new articles could be modeled as a sequence of appendices to their containing set. A feed of updates to a particular document could be modeled as just those meta-data elements concerned with updates. Are feeds used for anything other than lists of add and update events?