Data streams: new datapoints associated with identifiers

Examples:
Paperbuzz – Assessments of papers
Cobaltmetrics – Meta-identifier / ID-free

Tools are often built around one or two identifiers.
Streams are often built around one or two use cases (a supply of some sort of data format)

Elements:
Namespaces vs Identifiers vs Intrinsics (content/multihashes)

Identifier clusters (linking related elements)
UUID for each element
Name in namespace