Redux is a powerful programming paradigm, but one of the most common criticisms is that it requires too much boilerplate code. For a long time, I agreed with that criticism: why would someone want to do so much work to get basic functionality?
I struggled with this question until I realized that I was thinking about it the wrong way. Saying Redux introduces boilerplate code is sort of like saying C introduces boilerplate – it’s that way by design.
That’s when I finally understood Redux. Before long, I built a simple framework that used Redux and automated away the majority of the boilerplate code that I was reluctant to write.
Let’s use the canonical javascript example: we’re building a todo list. To do this, you need to do the following:
On the backend:
With this backend API set up, I’d then go back to my frontend codebase and
After repeating this process ~10 times, I realized that I didn’t really need so much fine-grained control. I didn’t need to call it id
in one place and _id
someplace else. And if I could just settle on a single way for defining attributes, I’d save myself a lot of time and my users a lot of bugs.
So I built a small library called Resource Creator that automated as much of this process as possible.
Use the createResource
helper to bootstrap boilerplate everything:
// store/todos.ts
import { createResource } from 'lib/resourceCreator'
export const { actions, reducer } = createResource({
resourceName: 'todo',
collectionName: 'todos',
})
By doing this, you’ll get the following:
On the backend, define your resource:
# routes.rb
resources :lists
Then, create a controller that inherits from GenericObjectsController
.
# todos_controller.rb
class TodosController < GenericObjectsController
def resource
Todo
end
def resource_params
params.require(:todo).permit(:id,
:title,
:description,
:done
)
end
end
And you’re done. Something that previously took an hour now takes 5 minutes.