Template:Tutorial Approach Comparison

Approach Comparison

Here is the architectural picture to demonstrate the interaction between Model, View, and Controller/ViewModel.

The main differences are that Controller changes to ViewModel and there is a binder in MVVM to synchronize data instead of a Controller.

MVC Advantage

• very intuitive, easy to understand
• control components in fine-grained

For those who use ZK for the first time or beginners, we suggest you using the MVC pattern. Because it's easy to use and debug.

MVVM Advantage

• suitable for design-by-contract programming
• loose coupling with View
• better reusability
• better testability
• better for responsive design

Both approaches can achieve many things in common, but there are still some differences between them. Each of two approaches has its strength. Building an application with MVC approach is more intuitive because you directly control what you see. Its strength is that you have total control of components so that you can create child components dynamically, control custom components, or do anything a component can do.

In MVVM, because ViewModel is loosely-coupled with View (it has no reference to components), one ViewModel may associate with multiple Views without modification. UI designers and programmers may work in parallel. If data and behavior do not change, a View's change doesn't cause ViewModel to be modified. In addition, as ViewModel is a POJO, it is easy to perform unit test on it without any special environment. That means ViewModel has better reusability, testability, and better resistance to View change.

To summarize, a comparison table is illustrated below:

	MVC	MVVM
Coupling with View	Loose with layout	Loose
Coupling with Component	Tight	Loose
Coding in View	Component ID	Data binding expression
Controller Implementation	Extends ZK's composer	a POJO
UI Data Access	Direct access	Automatic
Backend Data Access	Direct access	Direct access
UI Updating	Manipulate components	Automatic (@NotifyChange)
Component Controlling Granularity	Fine-grained	Normal
Performance	High	Normal