a word on preprocessor directives

Hey everyone,
I was typing out some code when a question popped up in my head that I was dying to have the answer to. Here goes: why is it that when you #include a header file of your own creation (what's the correct terminology? Send a header file to #include?) you write out code like the following:

Code:

#include "whatever.h"

but when dealing with files that come with your compiler like iostream, you write:

Code:

#include <iostream>

What exactly is the necessity for the difference in syntax...why arent you able do the following:

Code:

#include <whatever>

or

Code:

#include "iostream.h"

Just call this intense curiosity, any insight into this question would be more than appreciated. Thanks a lot-Chap

> #include "whatever.h"
Use this when you're including a file which is closely related (in the same directory) as the source file. foo.cpp would typically #include "foo.h".

> #include <whatever.h>
There are 2 cases
1. You've written a library you want to make use of in multiple projects
2. You've downloaded a library you want to make use of in your projects.
The <> tells the preprocessor to search all the standard include directories for the file. In most cases, extra files are not in the standard places, so you have to tell it where else to look.
In gcc, you would use

Code:

gcc -I/path/to/otherlib -I/path/to/mylib foo.cpp

People using IDE's should look at the project settings, and edit the "additional include directories" field.

> #include <whatever>
Is the same as the previous example when it comes to locating the file.
It may be that files without a .h imply that namespaces are being used, but that's only a guess on my part.
Certainly any well-written C++ library would use namespaces of some sort or another.