Create an Earley Parser written in the MUMPS (or M subset of Cache, or Ensemble) that is capable of parsing a Host File of text, which reads the Context Free Grammar data from one or more VA FileMan datastructures, and generates a parse tree which is a VA FileMan datastructures.
The context free grammar for parsing the language to use when parsing the text file will be capable of including an action (in MUMPS) to call when a rule is matched. Input to the rule action should include the parse-tree and the currently parsed text that matches the parse-tree, as well as text file information that is sufficient to allow the rule action to read the next text.
Finished work will be a KIDS file complete with Data Dictionaries and routines in the KBADG namespace. Result will be released under an open source license.