[Python-mode] speech recognition/Emacs match up

[Eric S. Johansson]

Many moons ago, I mentioned the need for some special features so that one can 
more easily create Python using speech recognition. I've got a fair start on 
some of the symbol manipulation techniques (http://blog.esjworks.com) but it 
counts on the ability to identify and isolate symbols based on their implied 
type signature.

 From the speaking perspective, there are only a few types of symbols. Class, 
method, instance, variable.  All of these are relatively difficult to detect in 
many circumstances unless a databases Of symbols and their type signature is 
maintained.

But if it isn't, all is not lost as long as I can fall back on the generic 
"symbol" so I can say things like "choose symbol 4" and have it select the 
fourth symbol in the line. Along the same vein, it would be nice to have the 
ability to choose an index, a full set of arguments or in individual argument. 
Super special would be being able to reduce scope after you select something 
like an argument which is a method with multiple arguments of its own. Being 
able to select that argument by name (choose argument 3) and be able to narrow 
your focus to just that argument has the potential for being very good for 
speech use and for focusing awareness.

How hard would it be to do these things?