Defender methods and compatibility

[Brian Goetz]

>> This is a trivial detail that's not worth holding back progress for
>
> Was somebody proposing to hold back progress?

The sad reality is that a fair portion of the postings to lambda-dev, no 
matter how well-intentioned, have a negative effect on progress.  (The 
exceptions are mostly the bug reports (thanks Thomas, keep 'em coming!))  It 
takes time to read, consider, and respond.  And messages nearly always arrive 
at a time when you are working on something else, causing context switch 
overhead.

The question is whether the potential incremental benefit that comes from the 
discussion is worth the incremental cost in time and schedule risk. 
Unfortunately nearly none of the contributors in this forum have all the facts 
(e.g., schedule, budget, and other non-public constraints) with which to make 
that tradeoff.  But every discussion has a real cost to the community in the 
form of incremental schedule risk.  I can tell you for a fact that the 
discussions on this list have frequently contributed to the schedule risk for 
delivering lambda.  So please, use some judgment.  It may be fun to argue 
language design, but excessive argument here hurts the entire Java community.

This specific discussion does not show signs of convergence, and as such has 
such an an impediment to progress.  At this point I see only downside in 
continuing the argument.  What has been said here may well be considered by 
the EG when the appropriate time comes.