The first subject for my reverse modeling is Twitter.

Higher requirements

There are three major types of actors in Twitter: user (unregistered or registered), advertiser, and administrator. Thus I organized higher requirements of this system by the types of user.

To limit the number of subsequent works, I omitted several functions from the scope (strikethrough).

• User

  • All users can:
    • create account if they meet the age restriction
    • view specific user’s
      • profile
      • “Tweets” tab
      • “Tweets & replies” tab
    • view specific tweet page
    • “You might like” card
    • What’s happening card
    • use “Explore”
      • search
        • user by
          • name
          • username
        • tweet by
          • keyword
          • hashtag
  • Registered user can log in
  • Logged-in user can:
    • visit “Home” to see Timeline
    • post new tweet
    • act on specific tweet:
      • reply
      • retweet
      • send like
      • report
    • act on specific user:
      • follow
      • add to list
      • mute
      • block
    • act on “Lists”
      • create
      • view
      • edit
      • delete
    • act on hashtag
      • create
      • use the existing one
      • view tweets that use the same one
    • view “Notification”
      • who mentioned them
      • who liked their tweets
      • who retweeted their tweets
    • act on “Messages”:
      • send messages to specific user
      • view received messages
    • act on “Profile”:
      • view
      • update
        • Name
        • Bio
        • Birth date
      • delete
    • act on account info
      • update
        • username
        • password
    • view specific user’s
      • “Media” tab
      • “Likes” tab

• Advertisers serve ads to users based on attributes such as

  • gender
  • language
  • geographic coverage
  • keywords
  • interests

• Administrator can activate/inactivate a specific registered user

Using the requirements listed above, I’ll create the initial domain model.

Nouns used in the higher requirements

They are the candidates of classes:

• Actor
  • User
    • Unregistered
    • Regestered
  • Advertiser
  • Administrator
• User attributes
  • gender
  • language
  • geographic coverage
  • keyword
  • interest
• Age restriction
• Profile
• Tweet
• Ads
• Timeline
• Like
• List
• Message

Domain model diagram created

After the initial domain modeling session, I’ve got a diagram below:

This diagram is incomplete, messed up, and may (must) even contain mistakes! But I will correct these problems during subsequent steps of ICONIX process.