Skip to content
Switch branches/tags

Latest commit


Git stats


Failed to load latest commit information.


Build Status


Public broadcasting in Germany receives €8,000,000,000 (eight billion euros) annually, yet it is subject to no public feedback, ranking, or even debate on what constitutes value or quality.


On you can say how your €17.50 per month should be spent. It is a proof of concept how digital democracy can work for publicly funded media and it is a win-win situation for both sides: More influence for the audience. More data for broadcasters.


Rundfunk mitbestimmen is maintained by the community. We have regular meetings, run online pair-programmings and tutorials in our online learner community at Agile Ventures. You can join our Slack here and then find us in our channel #rundfunk-mitbestimmen. Here is the Youtube Playlist of our recent meetings or pair-programmings:

Community pair-programming/meeting

Directory Layout

This repository contains three important folders:

  1. frontend (EmberJS)
  2. backend (Ruby on Rails)
  3. features (Cucumber/Capybara)


The backend is responsible to store the data. Who wants to pay for which broadcast and how much? Users are related to broadcasts via impressions in the database. The response on the impression model can be either negative, neutral and positive and indicates whether a user wants to give money to a broadcast. If the response is positive, the amount further specifies how much to pay for a broadcast. So, the sum of all amounts per user must not exceed the monthly fee of 17,50€ per month.

ER diagram


The frontend should be as easy to use as possible. The user can login and register, fetch a set of not yet voted broadcasts, decide about broadcasts, assign money and see the public statistics. Most of these actions will send requests to the backend. The frontend should be comfortable to use, e.g. by default amounts are evenly distributed with the option to set the amount explicitly.

Process diagram


We do full stack testing with Cucumber and Capybara. We specify the requirements as user stories in our Github issues and implement them as cucumber features. The cucumber features are a good starting for you to understand the current behaviour and the reasoning behind it.

Here is our model how to write tests. The cucumber tests are at the top. As they test the entire stack, cucumber tests tend to be rather slow in execution but in return they deliver some confidence that the system works. Testing pyramid

Installation and Usage with Docker

Make sure you have docker and docker-compose installed:

$ docker --version
Docker version 18.05.0-ce, build f150324782
$ docker-compose --version
docker-compose version 1.22.0, build unknown

Clone the repository:

git clone

You can setup the development environment with:

cd rundfunk-mitbestimmen
docker-compose up

This can take a while... As soon as this is finished, create the database and run migrations with:

docker-compose run --rm backend rails db:create db:migrate

App is running on localhost:4200

If you want, you can create some seed data

docker-compose run --rm backend rails db:seed

Run frontend tests:

docker-compose run --rm frontend ember test

Run backend tests:

docker-compose run --rm backend bin/rspec

For fullstack testing, use the provided docker-compose override:

docker-compose -f docker-compose.yml -f docker-compose.fullstack-testing.yml up

When all containers are up, run the cucumber tests in the fullstack service with:

docker-compose run --rm fullstack bundle exec cucumber

Local Installation

Make sure that you have a recent version of node, yarn, EmberJS, ruby, Redis and postgresql installed before you proceed. E.g. we have the following versions:

$ node --version
$ yarn --version
$ ember --version
ember-cli: 3.1.4
node: 10.4.1
os: linux x64

$ ruby --version
ruby 2.4.1p111 (2017-03-22 revision 58053) [x86_64-linux]
$ psql --version
psql (PostgreSQL) 9.6.5
$ redis-server -v
Redis server v=5.0.1

Clone the repository:

git clone

Install dependencies and run migrations:

  1. Install dependencies for full stack testing
cd rundfunk-mitbestimmen
  1. Install frontend dependencies
cd frontend
yarn install
  1. Install backend dependencies
cd ../backend
  1. Setup the database

(OPTIONAL): Customize the file backend/config/database.yml to match your local database configuration.

Now make sure that postgresql database is running

[linux]$ sudo systemctl start postgresql
[macos]$ brew services start postgresql

Create the databases and run the migrations:

rails db:create db:migrate
  1. If you want, you can create some seed data

You need to have redis running

[linux]$ sudo systemctl start redis
[macos]$ brew services start redis

Then you can seed the data

cd backend
rails db:seed
cd ..


Start the backend and sidekiq:

cd backend && rails s
cd backend && bundle exec sidekiq

Open another terminal and start the frontend:

cd frontend && ember server

App is running on localhost:4200

Full stack testing

Run the frontend server:

cd frontend && ember server --environment=fullstack

Open two more terminals and run the backend server and sidekiq:

cd backend && rails server --environment=fullstack
cd backend && bundle exec sidekiq

Open yet another terminal and run the tests:

bundle exec cucumber

These environments serve the purpose to

  1. Stub out external services in the frontend, e.g. authentication via Auth0.
  2. Use a separate test database in the backend, which will be cleaned after each test run.

If you want to run chrome or firefox instead of headless chrome, you can set an environment variable:

bundle exec cucumber DRIVER=chrome


bundle exec cucumber DRIVER=firefox

Frontend tests

cd frontend && ember test --server

Backend tests

cd backend && bin/rspec


See our detailed contribution guidelines 🔍

We use this project board as our central issue tracker. Issues are ordered by priority and you can filter for good first issue if you are interested in a beginner-friendly task. Additionally, issues are tagged with backend and frontend depending on where code needs to be changed.


Don't be afraid of our identity provider Auth0. In development environment your login will reach the "rundfunk-testing" instance of Auth0. This will not pollute the Auth0 instance used in production.

Workflow for Behaviour Driven Development with Cucumber:

  1. Fork it 🍴
  2. Pick an issue from the backlog
  3. Create your feature branch: git checkout -b [issue number]_my_new_feature_branch
  4. Createfeatures/[site of change]/your.feature and copy+paste the feature description from GitHub
  5. Boot both frontend and backend as described in the section about testing
  6. Run bundle exec cucumber
  7. Append the terminal output to features/step_definitions/steps.rb and write expectations
  8. Run bundle exec cucumber - tests should fail
  9. Implement the feature
  10. Run bundle exec cucumber - tests should pass
  11. Commit your changes: git commit -am 'Add some feature'
  12. Push to the branch: git push origin -u [issue number]_my_new_feature_branch
  13. Submit a pull request ❤️


Our build server Travis CI takes care about automatic deployment. Approximately 15 minutes after your pull request was merged into master, you should see the changes in production.

Useful Links


See the LICENSE file for license rights and limitations (MIT).


How should your broadcasting fees be spent? Publish your support and improve the program.





No releases published


No packages published