Crowdsourced trackers

[orthecreedence]

This is kind of a meta issue. Given that we have three main tracker types (occupations #89, resources #17, and processes #137) it makes sense to have some kind of abstracted process for crowdsourcing and creating these trackers.

I'm thinking some sort of tree approach makes sense. Each of these main types (occupation, resource, process) has a root type. From there, new entries can be created as children.

Right now, the rough working idea is this:

• Someone proposes a changeset (any number of changes) with some included rationale for the changes
• The change must get some level of confirmation ie "ok, yes, this isn't a braindead idea" from some percentage of the network agents
• Once confirmed, the change moves to a general vote

(See #60)

Standard object definition:

Translatable {
  translations: {
    en-US: {
      name: <english tracker name>
      description: <english tracker description>
    }
    <localized names/descriptions> ...
  }
}

Tracker: Translatable {
  id: <unique id>
  parent: <unique id>
  children: [ <unique id> ... ]
  <type-specific data> ...
}

Possible change operations:

• Insert new node. Can be leaf node (no children) or inserted between parent node and child (the new node is child to the original parent, and parent to original child)
• Move node, and children, to another location in tree
• Delete a node (all children will be adopted by deleted node's parent)
• Add translation
• Update translation
• Remove translation
• <type specific adjustments>

So you could have

occupations (1)

then applying

[
  {insert: 1234, parent: 1, translations: {en-US: 'doctor'}},
  {insert: 5678, parent: 1, translations: {en-US: 'nurse'}}
]

would result in

occupations (1)
  doctor (1234)
  nurse (5678)

which is fine, but

[
  {insert: 10, parent: 1, child: 1234, translations: {en-US: 'healthcare'}},
  {move: 5678, parent: 10},
]

might be a bit more tidy:

occupations (1)
  healthcare (10)
    doctor (1234)
    nurse (5678)