Exposing compound objects on the web via Resource Maps is only part of the solution; the Resource Maps and its referenced resources need to be discovered to really become part of the web graph. OAI-ORE proposes two complementing approaches with this regard:

• Harvest type discovery, which consists of making batches of Resource Maps available through existing mechanisms such as RSS, Sitemaps, and OAI-PMH.
• Linked Data ¹¹ type discovery, which uses HTTP headers received in response to dereferencing the URI of a component of a compound object to point at the Resource Map(s) that correspond(s) with the compound object. This is shown in Figure 8 where a crawler lands upon splash page S and is pointed at the Resource Map R by means of a HTTP LINK header contained in the response to an HTTP GET request issued against S. From R, the crawler can obtain a Resource Map representation, and hence a list of all resources that are part of the compound object, as well as further internal and external relationships.