Question

Is it a good idea to use self-invoking functions to hide functions and variables like this?

Is there a better way?

Is it slower than declaring functions and variables on the same level?

validator = require 'json-schema'

actionA = do ->
  schema =
    type : 'object'
    properties:
      param1 : { type : 'string', required : true }
      param2 : { type : 'string', required : true }

  encode = (results) -> JSON.parse(results)

  (params, callback) ->
    res = validator.validate params, schema
    value = if res.valid then encode(params) else false
    callback value

actionB = do ->
  schema =
    type : 'object'
    properties:
      param1 : { type : 'integer', required : true }
      param2 : { type : 'integer', optional : true }

  encode = (results) -> JSON.stringify(results)

  (params, callback) ->
    res = validator.validate params, schema
    value = if res.valid then encode(params) else false
    callback value

Thank you for your help.

Answer 1

Yes, I believe it’s fine. In fact, when you create “classes” in CoffeeScript

class Test
  constructor: ->
    @foo = 'bar'

it compiles to this:

var Test;

Test = (function() {

  function Test() {
    this.foo = 'bar';
  }

  return Test;

})();

Which primarily does the same thing as the code you posted.