blog-feedback-app

A small ClojureScript + Rum + DataScript app used as an example for a blog post, to write feedback to that blog post.

Idea

Attach a script to a post that make it possible for readers to write feedback to that post.

App design

The application is based on the structure of the posts of my blog:

  <div class='body'>
    <div class='section' id='section-title'>
      <h2>Section title</h2>
      <p>Text<p/>
    </div>
  </div>

At start the script looks for elements with class="section" and, for each of those elements, gets the id and title of the section and attach a feedback component to it.

Each component, attached to a post section, adds a feedback form to add new comments and a div element with a list of previous comments to which a reader can reply to.

Feedback data should be stored in a DataScript database and the components will react to changes of the database.

Database schema:

(def schema
  {:section/id {:db/unique :db.unique/identity}
   :comment/id {:db/unique :db.unique/identity}
   :comment/section {:db/valueType :db.type/ref
                     :db/cardinality :db.cardinality/one}
   :comment/parent {:db/valueType :db.type/ref
                    :db/cardinality :db.cardinality/one}})

How to run the application

This project includes an example post in resources/public/index.html to be used as a base during development.

To run the application you need the clj command line tool:

$ clj -A:fig:build

This will generate the js code and will open the default browser with the application with a Figwheel session and a REPL to interact with it.

At this point you are ready to write feedbacks in the page and interact with the application within the REPL session.

To create a release build run:

$ clj -A:fig:min

When the build will finish, the application file will be available at target/public/cljs-out/dev-main.js, ready to be embedded in your page.

Current state

Right now this is just a browser application and there is no backend but it is, more or less, ready by design to interact with a server side application. Changes to the database are sent as messages to a channel and a listener gets messages out of it and updates the database. It should not take much time to send events to a backend and to subscribe to a websocket to get realtime updates.

Next steps

Add a backend with:

• a durable storage
• an API to get previous messages
• an API to create new messages
• an API to delete messages
• a websocket for realtime communication
• optionally an authentication system

Enough for the README.md, more details in the upcoming post.