These splash pages have come to de facto represent the compound object “as a whole” on the web, and, as a result, a convention has emerged to use the URI of the splash page as the URI of the compound object itself. While this approach is useful for human users, it is problematic from a machine re-use perspective because:

• Machine interpretation of splash page information is difficult or impossible due to the lack of standards for its structure; and
• According to the Web Architecture, the URI of the splash page merely identifies the resource that is the splash page, which is actually only a component of the compound object and not the compound object itself.

In addition to these problems related to identifying the compound object and defining its structure, the web architecture has no method to explicitly reference a resource within the context of a compound object. This functionality is important for scholarly communication because an existing resource can be re-used as a component of any number of compound objects. For example, a specific cellular image may be part of one image web illustrating confocal microscopy techniques and of another concerned with cancer therapy. For provenance tracking and citation, it is important to have the ability to reference a resource as it exists as a component of one or other specific compound object (e.g., the cellular image as a part of the cancer therapy image web), because the exact meaning of the resource or of a reference to it can be dependent on the context provided by such object.

The goal of the ORE work is to address this shortcoming and define an infrastructure layer over the web architecture allowing interoperable use and re-use of and reference to compound information objects across a variety of networked applications. This ORE layer explicitly does not replace or redefine any core web architecture concepts. Indeed, it leverages them fully as part of the solution to the problem of expressing compound objects.

The ORE layer expresses the boundaries of a compound object in a manner that can be processed by machines and agents. This can be viewed as the specification of a machine-readable splash page that lists the components of a compound object, as well as their internal and, optionally, external relationships. Such a specification would support:

• Re-use of a compound object and its components across web-savvy applications.
• Reference to a compound object and its components in a manner that supports an understanding of their “compound object context.” This would allow reference to a resource as it exists in the context of a specific compound object, distinguishing that from a reference to the same resource as it exists in the context of another compound object, or as just a resource in its own right.
• Machine discovery of compound object information.