Think of any resource on the Internet, be it a document, a web site, an application, a database or whatnot. Sooner or later, said resource is updated (or outdated). That update is, in effect, an event applicable to said resource. Currently the parties interested in following such events are required
to observe a heterogeneous amalgam of secondary resources: RSS feeds, e-mail newsletters, SMS notifications, interpreting various free or proprietary data streams (Windows Updates, Adobe updates, Apple updates, Nokia updates, yum, apt-get, etc.)
This idea calls for a universal format for resource updates, possibly using a structure based on IETF's RFC 5785 "well-known URIs". As such, the URI proper would have to be persistent within the actual resource content (e.g. the original URL of an HTML file would have to be available in a predictable manner in the actual HTML file; the same would have to be applicable for any resource, such as PDF documents, Windows executable files, Linux packages and so on). Furthermore, the standard would have to be designed in such a way as to enable an automated agent to determine the URI for the resource's associated Universal Resource Update (URU) using simple canonization on the resource URI and the eventual URU well-known location (for instance the Google Chrome installer would be known to have originated from domain google.com, and the well-known location for URUs could be /default.uru; in this case, an automated agent could determine that any updates relevant to Google Chrome can be found at http://google.com/default.uru).
The URU file itself would be an XML file containing the latest version of each relevant resource available on the domain hosting the URU file, plus whatever information deemed necessary (name, author, license, last version, release date, URI for historical updates, dependencies, etc).