Application structure

A full-stack LoopBack application follows the model-view-controller application pattern.

• Models are LoopBack models, defined in the application's /models directory.
• Views are defined in the application's /web/views and /html5/views directories. Why are there two views directories? What is the diff?
• Controllers are defined in /html5/controllers.

A full-stack LoopBack application has four sub-directories:

• The /api directory contains code for a "standard LoopBack" app.
  • app.api.js - main app file
  • configure.js - configuration settings, for example hostname and port, database logins, and so on.
  • package.json
• The /models directory contains "standard LoopBack" model definitions.
  • index.js - Defines each model using the JS files in this directory, for example: exports.Todo = require('./todo');
  • One JS file to define each model with loopback.DataModel.extend(); for example todo.js.
  • package.json - (Not sure why this is needed here?)
• The /html5 directory contains:
  • app.html5.js - main file
  • configure.js - gulp build file (?)
  • package.json - (Not sure why this is needed here?)
  • /controllers directory
    • app.ctrl.js - main application controller
    • One .ctrl.js file for each model; for example, user.ctrl.js.
  • /views directory - contains .html files, including one for each model; for example, user.html.
  • /build directory - output of build process? Is this gulp output?
  • /bower_components directory contains files that Bower uses to manage dependencies for front-end packages.
• The /web directory contains: ** app.js - defines routes (what else?) ** run.js - Used to run the application (?) ** /views directory - contains template files, such as .ejs. ** package.json

NOTES: For simplicity, I did not include test directories. These can be described in an addendum.

Questions

Why are there three package.json files? Could it be done with one?