Re: proposal for call_ and str_

Mark Hahn wrote:

> If that "proxy lookup" fails to find the attribute, then the object
> itself will be searched for the attribute as if the proxy object
> didn't exist.

I put in this provision because I knew it solved some problems that I was
thinking of.  What I didn't realize was that it solved those problems
because it now isn't a simple dual namespace where each object is a
different namespace.  Attributes of the main object are in both namespaces
unless overriden by same-named attributes in the proxy object.  Only when
overriden are there objects in seperate namespaces.

Even though this isn't the obvious simple dual namespace solution, I think
it gives you all the benefits you want.  In other words when you want to
have two methods with the same name you can.

Another way of looking at it is any method in the main object (instance)
namespace is automatically copied into the proxy (factory/class) namespace
if it isn't overriden.

This also solves the directed method calling problem I've been rambling on
about in all the proposals.  I fixed it by accident.  Yeah me :-)