Java Swing - Should JTree be used with TreeModel - MVC design pattern

Question

I am going to use JTree in my Java Swing Desktop Application. I read about MVC Design pattern and according to it, we should use Model classes of Swing Components to separate the datamodel of a JComponent from its presentation part.

So my quick question is as follows:

• JTree(TreeModel newModel)

• JTree(Object[] value)

• JTree(TreeNode root)

• JTree(Vector<?> value)

• JTree(HashTable<?, ?> value)

Out of the above options for creating a JTree,

Q1. Is it always better to go with JTree(TreeModel newModel) to apply the MVC pattern?

Answer 1

It depends upon your needs. I tend to think Swing components constructors taking as input non-model objects are in fact shortcuts for fast prototyping. indeed, behind the hoods, the Swing component will create a model from the input object, since it requires a model object to have all the events correctly sent.

As a consequence, here is my advice :

• For a fast prototype, you may consider the use of these "mock-like" constructors
• For a "real-world" application, don't even think about them, since sooner or latter will arise the need for specific event sending (to change one node rendering, making the tree grow, ...)

Answer 2

As Riduidel said, JTree always uses a TreeModel internally, so the other constructors are really just for convenience.

There's also the setModel(JTree) method, which will set (and replace) the model. In a non-trivial application you'll probably want to build the frame and it's components before filling in the data.

I don't consider any of the constructors as being non-MVC. What's more important in that regard is that you keep code responsible for the data, UI, and logic as separate and non-dependent as possible. That allows you to better unit test your code, and it helps with flexibility and re-usability.