Why the one request bean?

Question

A general question: in all JSF books and examples they always use specialized managed request beans for each business process/flow, etc.
In JSC, there is this one huge monolithic RequestBean1 where everyone tacks on extra objects and properties.
if this entire bean is getting instantiated every time any sort of JSF request happens, wouldn't it get very "heavy" after a while? Especially, in terms of memory usage...
Why does JSC implement request managed bean as one big bean instead of allowing multiple, smaller, fine-grained objects?

Answer 1

OK, then...but if we're supposed to use our own managed beans (as it should be) why does every page return a reference to RequestBean1 anyway via getRequestBean1()?
protected RequestBean1 getRequestBean1() {
return (RequestBean1)getBean("RequestBean1");
Doesn't this code encourage incorrect programming practices by implying everyone should be piling every possible request property into the same and one bean?