[JSR308] evaluating Java language changes

[Eugene Kuleshov]

  It seems like we already have several proposals for the Java language 
grammar changes (and we may see more of them). So, I wonder how should 
we evaluate each of those proposals? What characteristics should be 
evaluated for each proposal? Probably the most obvious metrics are 
obviousness, unambiguity and verbosity. What other metrics we can take 
into the consideration?

  Michael Ernst also suggested to actually use the current prototype, 
but this evaluation is tough because it is bound to specific use cases 
and there isn't generic enough showcase that would allow to try most of 
the grammar changes.

  Any other thoughts?

  regards,
  Eugene