On Jan 27, 2008, at 11:40 PM, Gary Wright wrote:

> I think that any sequence of bytes, including the empty sequence  
> should be mappable to a symbol.  To me, the the Symbol class is a  
> guaranteed one-to-one mapping from a string instance to object.

I don't necessarily disagree, but I'd like to explore.

1. What general problem does this solve?

2. Does this idea of string mapping jive with the fact that  
currently  :+ is the same object as  :"+" is the same object as :"+ 
(binary)" ?  The latter two seem troublesome to me--there's an  
abstraction leaking out here.


Dave