h1ghlevelb1ts

I am reading a book on JavaServer Faces (The Complete Reference.....) since I am supposed to teach the subject next week at a course on Java and web programming. One code example got me started on one of my favorite subjects - the abundance of null checks in source code. This is the example code I saw inside a custom converter example class:

public String getAsString(FacesContext context,
    UIComponent component,
    Object value)
{
    if( context == null || component == null )
    {
        throw new NullPointerException();
    }
    // and obviously some more code here....
}

This doesn't make sense at all since if the params are null there will eventuelly be a NullPointerException raised anyway. This method is supposed to be called by the JSF framework so it makes even less sense. Why would a framework call a custom converter class with null values? I saw similar code yesterday in the huge code pile I am working with now for a customer. The difference was that a checked exception was thrown instead of an unchecked. That is even worse since every caller of the method need to deal with that exception and - in this case - the source fault (null value of a param) was not mentioned at all. Just a generic "this was not good" message. I suggest that null values can take care of themselves. If someone passes in null for a param that never should be null a NullPointerException is a brilliant outcome with the stack trace showing us exactly where the null value was discovered.