TODO.txt

Word Processor Architecture:
 - DOM tree structure
 - Traversal iterates over each node in the document linked list
   - Checks for presence of children, then goes to next sibling
   - Recursive? Or depth-first-search technically
 - Nodes index into a separate format piece table?
   - Layout and style information is stored here
   - This allows us to select subsets of the piece table to view
     and still have formatting propagate
   - Still supports undo/redo
 - Designed primarily for 'fragments'
   - Can view sections in any order and have redo/undo work appropriately

Container Nodes:
 - List
 - Table
 - Equation

Goals:
 - Support most of what google docs does
 - All pango attributes are usable
 - I18n and bidirectional support (at a later date)
 - Tables/Images/Lists/Equations
 - Full Undo/Redo
 - Fast, Lightweight, Native
   -> Entirely in C (no webkit)
 - Eventually split out into a reusable Rich Text widget
 - Save to ODT (maybe also Markdown/HTML?)
 - MPL-2.0 licenced. Somewhat permissive, still copyleft.
   -> Compatibility with LibreOffice source code

Non-Goals:
 - Compete with Microsoft Word
 - DOCX/OOXML
 - Forms, Macros, Scripting
 - Any sort of advanced layout
   -> Use Publisher or Scribus instead