Our first pass plan of approach:

1. Load the full document into a DOM
2. Unwrap all parsoid-generated section tags except the requested one from the URL parameter
3. Convert the full document into the VE data model
4. Build VE view tree for the selected DM node object

Current assumptions:

• We want to focus on sections as currently defined by Parsoid (not paragraphs/sentences as sections)
• The "selection" (document model node) loaded will be a single node rather than a group of siblings

Limitations of approach:

• Changing from one node to another will be more difficult

Benefits of approach:

• Don’t have to deal with potential corruption of the document tree hierarchy
• Avoid the addition of unwanted section tags