Class overhead, function overhead and parameter overhead

Question

Hi all,
I am interested in class overhead, function overhead (each function in a class) and parameter overhead (each parameter in function).
I saw somewhere that class overhead are 200 bytes.
Is it wise to give up on class and go for functional programming? (Only classes required to run the j2me app are created and nothing else)
How does function get called? Does it work like a stack? Where each parameters get pushed into the stack and later on pop out when control pass on to the function?
If your program is full of functions doesn't this make it very processor intensive?
What is the best solution?
Thanks in advance!

Answer 1

Hallo,
silly question: but what exactly do you mean when you say "full of attributes and methods that will never be used for this instance"?
I assume that when you have created your class you thought about the attributes and methods you wanted to include, and have not included any unnecessary baggage. I can only think that, perhaps, you have a class X that you would like to reuse, because it does all that you want. The only problem is that, for your specific problem, it carries a lot of excess baggage with it. You only need a few of the attributes and methods. Am I correct?
Such a class reuse will certainly incur some overheads, but let us try to analyse where they are:
1. Extra Code. The 'big' class is already defined and probably already loaded, so there should not be any overheads for the code.
2. Extra Data. Certainly creating a new instance of the 'big' class will require more memory. The question is how big is 'big' class, and how many instances do you want to create or have in existance at any one time? If the class is less than a few kB big and you only need a few instances at any one time, then you can afford to ignore the overhead.
3. Performance. Every time that you create an instance of the 'big' class, the processor must do some work to initialize it. The question is, what is required to initialize the class? This depends on what is in the constructor.
4. Programmer Time. If there is an existing class, then it will (probably, depending on documentation) take less time to reuse the class
than to write another, streamlined class. Someone else will already (hopefully) have debugged the 'big' class.
I think my view would be to reuse 'big' class unless there good grounds not to do so (see above). Your time is more important than a few overheads.