The necessary steps to get the application up and running.
-
System dependencies
docker
docker-compose -
Configuration
docker volume create --name=chatbundle
docker volume create --name=chatdata -
Application execution
docker-compose up -d --build -
Database creation
docker-compose exec chat bundle exec rails db:create -
Database initialization
docker-compose exec chat bundle exec rails db:setup -
Node initialization
docker-compose exec chat yarn install -
How to run the test suite
docker-compose exec chat bundle exec rails test -
How to access application
GET http://localhost:3000/ -
Credentials
one@example.com
password
The features were selected based on showcasing an application that runs on a full stack and has a pleasant, familiar, look-and-feel. To demonstrate the front end features, specific back end features were selected to ensure the UI has the data it requires. Due to the front and back end interdependence, more exotic features such as gif suggestions and search were skipped as I feel these are biased towards more specialized front and back end development.
- See list of channels
- See history of messages in channel
- Persist chat messages in joined channels
- See list of channels
- See list of other users
With more time all of the application's features would be completed and verified via integration testing, augmented with unit tests and insuring proper code coverage. The front end refactored into component files. Websocket based channels implemented to facilitate a proper interactive experience. Jobs created for the bulk of model interactions to decouple them from the syncronious request/response cycle.
UX drives implementation. A familiar, intuitive interface is fundamental for the success of an application, else its features are not going to be used regardless of how useful they may be. Therefore a greater focus was given to presentation on the front end. For the back end the main assumption is that data that is supplied to the front end must be correct and verified to be correct. Therfore focus on the back end leaned more towards testing the API end points.