The view layer (An iOS architecture part II)

This is the continuation of the following post, the aim of the view layer (an iOs Architecture part II), is just to explain in detail how view layer will be built.

The view layer

When you were working on development tasks related purely related with view just shift your mind to the following schema:

The schema is almost MVP, almost because in some concrete cases does not fit (we will see later on). But the main components are views and presenters, and its role is practically the same


Views are the classes responsible of drawing the information in the screen of device. The input of this class is the model to printout and the output the events that view can not handle.

The events can be handled via:

  • Delegate protocol.
  • Callbacks.

In the implementation project I have used callbacks.

For instance if a view hast to print out a list of persons, as soon as view receives an array of persons will print out the list. When user taps on one of the persons, view catches the event, but does not know what to do with it, so bypasses the event to its parent presenter.

Composite views 🙄. Is quite common that a view will be complex enough to be splited in subviews. In the previous example, figure it out that now we want to print out also a header and a footer on the persons lists. The view will have access to its subviews, and as soon as receives the model from the presenter it will bypass also the models to its subviews. When one of its subviews will receives a user action, depending on the situation, will handle it or will bypass it to its parent presenter.

Heterogeneous Composite Views😣. I will start directly with the example, figure it out the home view of a very famous application like Spotify, LinkedIn, Facebook… In the same screen is presented several sections each one presenting information not related with the other sections.

Could your tell the number of lines for this presenter? Who would be the super-hero in case that some change would be required in the home screen? This is a clear smelt that MVP is not enough for such implementation🤔.

In case of such kind of complex views or in case that we want to reuse the same section views in two (or more) different screens the pattern that we use is MVPP (or MVP+). The idea is as simple as a view will be the container of another presenter (with its own views).

Every (sub)presenter will fetch data from business and will pass it to its views, this will help a lot in reducing the size of the main (super)presenter. Another great advantage of this pattern is that allows to reuse the presenter-view in some other screen.



Presenter is the abstract representation of a screen app, it has direct access to its views. Its role is quite clear fetch/pass data from/to Business (Domain and Data) layer on one hand and in the other transfer information between views and handle the events reported by views.

For a presenter its views are black boxes, at no moment the presenter knows if what is being drawn is a tableview, a collection view, a pageviewcontroller…. It transfer Models  to view, and wait for the events that view can not handle.

As well as in views, there are some events that even a presenter can not handle, for instance if user presses a button for dismissing the app screen view. In such case the presenter will bypass the event to its coordinator, … but this will be explained in Domain layer (part III).


The Presenter will be implemented as a subclass of a UIViewController, whilst for View we have two alternatives UIView or UIViewController. I prefer to use UIViewController, mostly because is easier to control the lifecycle of the view. But If you do not have to, there’s no problem in using UIView.

The project app just present two screens, and heterogeneous composite view and a simple view.

Home screen is a presenter of presenters and the second one is just a presenter. As you can see popular presenter has been reused in the home, this is why is so powerful this model presenter plus protocol.

Simple presenter (Popular).

The presenter is subclassed from UIViewController this allows us to represent it in the storyboard. The view could be subclassed from:

  • UIView. UIViews is the class implementation for handling views, so why not…
  • UIViewController. With UIViewController we have the extra that we can handle the view lifecycle much better than with a UIView.

In the story board the view is embedded in the presenter with a container view:

This is the implementation for the presenter:

In the prepare segue is where is handled the event that is coming from the view, and is bypassed to UseCase (… but this is domain part II).

🚫View could also call the UseCase and the compiler did not complain, but this is totally forbidden in this architecture🚫. Remember 3-tier layer philosophy.

And the view simply:

Print out the model and send back the button event.

Composite presenter (Home).

The Home presenter is a presenter of presenters, the unique case that has to do the superpersenter is just show/hide the sub presenters. But in the case of the implementation both are presented. The storyboard shows pretty clear the super-presenter (at the top), its container view that will hold the subcontainers for placing the presenter for the latest popular place and the most popular place.

The home presenter implementation:

And its view container:


In this post I have presented how to structure view layer using presenters and views. Finally the idea has been supported using a example project, the source code can be found here.

What’s next…

View layer is not finished yet, in the next post I will present how to handle font types (and its sizes), colours, accessibility labels (for automated testing) and localised texts (for translations).