Question

My setup: Rails 2.3.10, Ruby 1.8.7

I would like some feedback on where it is best to put logic code, and REST API, that touches multiple models in a single transaction. For example, a user needs to buy a product, it’ll involve

1. Checks if he has sufficient money (user model)
2. Check if the product is available (product model)
3. Calculate shipping charge (zipcode, product models)
4. Subtract money
5. Update product availability count
6. …

You get the general idea. Let’s say I need to provide a buy REST API, which controller should it go in? And where should the actual logic go? Should it be in the model associated with the controller? Appreciate any insights.

Answer 1

I don’t know what the convention is (if there is one), but I tend to do my multi-model transactions in a “noun-verb” format. For example, if a User wanted to purchase a Product, I would do:

class User < ActiveRecord::Base
  ...
  def purchase(product)
    product.logic
    self.step_3_profit
    etc
  end
end

The controller works similarly, although I usually think about it in passive voice (e.g., “what verb is being done to what noun”, rather than “what noun is doing what verb”. For example, if a Product was being purchased, it might be a POST to /products/1/purchase, with the following controller code:

class ProductsController < ApplicationController
  include SessionsHelper

  def purchase
    current_user.purchase Product.find(params[:id])
    do_view_stuff
  end
end

Using these “conventions”, I can easily locate the logic in my app by thinking about what verb is being done to what noun.