1) Scientists publish to both keep a record of science (record keeping) and facilitate the transfer of knowledge. However, these two goals may have different requirements and what satisfies one may not satisfy the other.
2) Nanopublications, which are assertions with attributes, can be used both as citations and to represent relationships between entities. They contain metadata about authorship, provenance, and evidence.
3) Representing scientific knowledge as interconnected nanopublications allows for reasoning across assertions and could provide an overview of a field while still allowing access to detail. This may help address the challenge of managing the increasing volume of scientific literature.