(2010/12/13 9:06), Ryan Davis wrote:
> Providing a standard AST definition for consumers != defining the internals. 

Definitely.  I'm not against a standardised AST itself; just in doubt for that
standard to touch MRI internals.  To make a live AST access possible more or
less new API is mandatory.  The sadness is almost every new feature added to
MRI are seen as Ruby standard these days; we do not want that.