From REST in Practice: Hypermedia and Systems Architecture:
For computer-to-computer interactions, we advertise protocol information by embedding links in representations, much as we do with the human Web. To describe a link's purpose, we annotate it. Annotations indicate what the linked resource means to the current resource: “status of your coffee order” “payment” and so on. We call such annotated links hypermedia controls, reflecting their enhanced capabilities over raw URIs.
... link relation values, which describe the roles of linked resources ... Link relation values help consumers understand why they might want to activate a hypermedia control. They do so by indicating the role of the linked resource in the context of the current representation.
I interpret the above quotes as saying that Hypermedia control contains both a link to a resource and an annotation describing the role of linked resource in the context of the current representation. And we call this annotation ( which describes the role of linked resource ) a link relation value.
Is my assumption correct or does the term link relation value actually describe something different?