Abstract : The naive documentary model behind the Web (a single HTML Web page retrieved by a client from a server) soon appeared too narrow to encompass all cto account for dynamic pages, content negotiation, Web applications, etc. The Semantic Web raised another issue: how could we refer to things outside of the Web? Roy Fielding's REST style of architecture solved both problems by providing the Web its post-hoc "theory", making it a resource-oriented application. Recent evolutions (AJAX, HTML5, Linked Data, etc.) and envisioned evolutions (Web of devices, ubiquitous Web, etc.) require a new take on this style of architecture. At the core of the Web architecture and acting as a unifying concept beneath all its facets we find the notion of resource. The introduction of resources was very much needed for the Web to remain coherent; we now have to thoroughly redefine them to espouse its evolutions through time and usages. From the definition and the characterization of resources depends our abilities to efficiently leverage them: identify, publish, find, filter, combine, customize them, augment their affordance, etc.