Question 1:
Unfortunately the one in which you you say is most intuitive is the least used in Android. As I understand, you should separate your UI (XML) and computational functionality (Java Class Files). It also makes for easier debugging. It is actually a lot easier to read this way and think about Android imo.

Question 2:
I believe the two mainly used are #2 and #3. I will use a Button clickButton as an example.

2

is in the form of an anonymous class.

Button clickButton = (Button) findViewById(R.id.clickButton);
clickButton.setOnClickListener( new OnClickListener() {
            
            @Override
            public void onClick(View v) {
                // TODO Auto-generated method stub
                ***Do what you want with the click here***
            }
        });

This is my favorite as it has the onClick method right next to where the button variable was set with the findViewById. It seems very neat and tidy that everything that deals with this clickButton Button View is located here.

A con that my coworker comments, is that imagine you have many views that need onclick listener. You can see that your onCreate will get very long in length. So that why he likes to use:

3

Say you have, 5 clickButtons:

Make sure your Activity/Fragment implement OnClickListener

// in OnCreate

Button mClickButton1 = (Button)findViewById(R.id.clickButton1);
mClickButton1.setOnClickListener(this);
Button mClickButton2 = (Button)findViewById(R.id.clickButton2);
mClickButton2.setOnClickListener(this);
Button mClickButton3 = (Button)findViewById(R.id.clickButton3);
mClickButton3.setOnClickListener(this);
Button mClickButton4 = (Button)findViewById(R.id.clickButton4);
mClickButton4.setOnClickListener(this);
Button mClickButton5 = (Button)findViewById(R.id.clickButton5);
mClickButton5.setOnClickListener(this);


// somewhere else in your code

public void onClick(View v) {
    switch (v.getId()) {
        case  R.id.clickButton1: {
            // do something for button 1 click
            break;
        }

        case R.id.clickButton2: {
            // do something for button 2 click
            break;
        }

        //.... etc
    }
}

This way as my coworker explains is neater in his eyes, as all the onClick computation is handled in one place and not crowding the onCreate method. But the downside I see is, that the:

1. views themselves,
2. and any other object that might be located in onCreate used by the onClick method will have to be made into a field.

Let me know if you would like more information. I didn’t answer your question fully because it is a pretty long question. And if I find some sites I will expand my answer, right now I’m just giving some experience.