The result and the concept just stays in your mind and it is you who created the source and latter you can be confident of what you say when you explain the same to others.

I hope you don't tell Maha I didn't ask you . Come back if you have anything more to clarify.

public class RAF
{
public static void main(String args[])throws IOException
{
long len =0;
try
{
RandomAccessFile myRAF = new RandomAccessFile("RAF.txt","rw");
System.out.println("Length of file is :" + myRAF.length());
}
catch (IOException ioe)
{
System.out.println("IOExceotion occurred.");
}
}
}

/* Missed you again by a few minutes Maha Anna. Well, that was what I was doing, and after the compilation and successful execution, a RAF.txt was created. So I guess it is true. Bit its contrary to what RHE states. Am I correct ? */

If you haven't you have a lots of correction to do.
The online R&H Errata linke is here.

regds
maha anna

Just an observation ...

The main() in the above code uses the throws construct
to specify the IOException and the method also catches it.
Though redundant, java compiler doesnot complain.

I din't know that ....

Regds.

- satya

Mainly this concept is used where as a designer you just design the skeleton of the overall s/w. (i.e) you are desigining

1. What are all the packages
2. what classes go under which package
3. what methods are inside which class
4. What Exceptions you expect the methods to throw

So the point which you noticed here comes into picture in the 4th point I stated above. You can just define all abstract methods and say, Fello collegues! These are the exceptions I expect these methods to throw. So while implementing the methods you take care of that. like that.

regds
maha anna