Factor out hash-map functionality

We have several collections including those for field and I/O context which make use of a hash-table to implement mapping functionality. At the moment this code is duplicated wherever it is needed.

While implementing a truly generic solution which can key on any data is prohibitively complicated at the moment a solution which keys on strings is within grasp and would reduce code duplication and ease re-use.

Locate the duplicated hash-table implementations used by the field and I/O context collections and compare their mapping interfaces. The work is done when a reusable string-keyed solution replaces the duplicated functionality while preserving the existing collection behavior.

Written by the indexing model from the issue text.