What kinds of changes need a JEP?

[mark.reinhold at oracle.com]

2017/11/7 4:38:02 -0800, maurizio.cimadamore at oracle.com:
> Many thumbs up on this Mark! Thanks for taking the time to put together 
> this revised proposal.
> 
> As a langtools hacker, I've often felt the pain associated with JEP 
> processes for what, in my mental model, was non-JEP work - an example of 
> such non-JEPs I came across over the last few years in langtools-land 
> are, IMHO:
> 
> http://openjdk.java.net/jeps/215
> 
> http://openjdk.java.net/jeps/217
> 
> I believe this work would classify now as regular javac enhancements and 
> would, therefore not be subject to the JEP process, which IMHO is good 
> news (not because I hate the JEP process, but because I hate abusing it 
> :-)).

These two JEPs, along with JEP 304 (GC Interface) which Roman mentions
nearby, strike me as interesting edge cases.

The text in each of these JEPs describes the motivation, background, and
design of each change.  Wouldn't that be useful to other developers who
work on the code, both present and future, even though these changes are
strictly internal?

The other typical places for such information are wiki pages, long e-mail
messages, and JBS issues.  These are harder to find, and in some cases far
less permanent, than JEPs.

Perhaps another criterion should be:

  - It has a non-trivial design

(although maybe that's just another sense of "significant").

What do you think?

- Mark