Unofficial wish list status.(Jul 2008)

[Russell Lewis]

Don wrote:
> I'm not sure why you wouldn't just make two classes, and apply invariant 
> at the class definition level.

As I understand it, the difference is that you can only call invariant 
functions on an invariant object.  So, if you have a reference to the 
class, and the reference is marked invariant, then all of the 
non-invariant functions can't be called.

Why would you have an invariant object, you say?  Well, a basic example 
is where you create an object, initialize it, then cast it to invariant. 
    (Presumably, you can guarantee, that it is invariant from the moment 
of the cast until the program dies.)