2IRR00Software Design

2IRR00 · Topic 11

Architecture: MVC & Observer

Model-View-Controller and the Observer pattern (push vs pull)

Model-View-Controller

MVC is an architectural pattern for GUI software that separates the Model (data + business logic), the View (presentation/IO) and the Controller (links view and model) — it says WHAT to separate, not HOW.

Model: internal representation of data; manages data, logic, rules, and requests on state.

View: a user interface that displays the model and accepts input. There can be several views of the same model.

Controller: links model and view; translates inputs into commands for the model/view.

MVC makes presentation independent of data, allows composing/swapping views (even at runtime), and separates concerns.

MVC can incorporate other patterns: Command (pass info), Decorator (add to a view), Facade (interfaces), Factory (create controllers), Observer (update views on model changes), Strategy (choose behaviour).

Double-click this block to add a note to Model-View-Controller.

Common mistakes

  • Saying MVC 'must' use the factory pattern (false) or that the MODEL manages view↔controller interactions (false — it's the controller that links them).
Double-click this block to add a note to Common mistakes.

Exam tips

  • Correct MVC MCQ: it separates model/view/controller but not HOW ✔; it can use the decorator to modify views ✔. Wrong: must use factory; model manages view↔controller.
Double-click this block to add a note to Exam tips.
Double-click this block to add a note to Memory aid.

Model-View-Controller practice

1 questions

Question 1/1 · 0 reviews recorded
Self-review
What does MVC define and what does it deliberately leave open?
Double-click this block to add a note to Model-View-Controller practice.

Observer (push vs pull)

Observer defines a one-to-many dependency: when the subject changes, all registered observers are notified and updated automatically, with loose coupling.

Subject: addObserver / removeObserver / notifyObservers + holds a list of observers. Observer: update().

Push method: the subject SENDS the (necessary) data in update(). Subjects must know what observers need → more coupling, often too much data sent.

Pull method: the subject only NOTIFIES that something changed; each observer PULLS the data it needs via the subject's getters. Less coupling, more independent/reusable, but can be less efficient (extra calls; observers must figure out what changed).

In both cases observers need some knowledge of the subject; coupling is reduced if they know it via an abstract class/interface.

Used in MVC/GUIs to update views, distributed event handling, notifications.

Double-click this block to add a note to Observer (push vs pull).
Push: subject sends itself
public void notifyObservers() {
  for (Observer o : observers) o.update(this);  // push (sends data)
}
// Pull alternative: o.update();  then observer calls subject.getState()
Double-click this block to add a note to observer-example.

Common mistakes

  • Describing pull as 'subject sends required data' — that's push. Pull = notify only, observer fetches.
  • Forgetting to trigger notifyObservers() after a state change.
Double-click this block to add a note to Common mistakes.

Exam tips

  • Pull (3-point open Q): subject only notifies; each observer decides what data to retrieve via getters → lower coupling, possibly less efficient.
  • Push increases coupling and often sends too much data.
Double-click this block to add a note to Exam tips.
Double-click this block to add a note to Memory aid.

Observer (push vs pull) practice

1 questions

Question 1/1 · 0 reviews recorded
Self-review
Explain the pull method of Observer.
Double-click this block to add a note to Observer (push vs pull) practice.