In this paper, we propose that the present state of Mathematical Knowledge Management (MKM) derives from two main imperatives: the desire to organize and encapsulate mathematical knowledge after it is produced and the desire to encapsulate the act of production. While progress in digital network technology has facilitated a confluence of these efforts, their original separation imposes an approximate rubric on MKM which may be used to help define the challenges facing the field. We propose that one of the main challenges lies in the question of fixed versus flexible ontologies and the related question of ontology resolution between applications. Finally, we describe Emkara, an application which adopts a flexible metadata definition in the archiving and retrieval of digital mathematical exchanges, queries, and grey literature.