TODO.md

1. simplify rng document
2. parse rng document into intermediate format (analogous to the xmlStructure object produced by the dtd parser)
3. convert intermediate format into xema format, as that seems to be what drives the entire application
4. store xema and see if it works in the current implementation
5. augment the xema, docspec, xematron and editor to add additional required features such as separate element names.

• add ability to separate element name from definition name and add reverse lookup during document parse
• preserve warnings present in the rng schema
• improve choices in the rng schema, preserve groups under choices?
• parse regex patterns for validation
  • actually validate using the regex patterns