Features for a possible 2.0 version:

• improve Programming API for writing schema (e.g. improve array access, indirect modification)
• improve reading performance by using a different approach than current, which is using recursion a lot