>>>>> Eli Zaretskii <eliz@gnu.org> writes:

> As long as the removal is not done by a program, but by humans, I think
> considering each candidate before producing the list is a reasonable
> compromise.

Right. I imagine that even without such a list, the process will look
something like this:

  1. Core developer looks at bug.
  2. Sees it is in "obsolete" code.
  3. Sees it has been "obsolete" for a while.
  4. Asks list, "Hey, can I just delete this code and the bug?"
  5. Hears deafening applause of consent.
  6. Happiness is restored.

Our biggest complaint about obsolete code is not its existence, but having to
tend to it; and so, this new policy gives us a way to do decisively deal with
it after a certain period of time.

Likewise, if an obsolete file never has any bugs, I'm not opposed to letting
it stay in "obsolete" forever. And yet, if someone requests that it be removed
after the appropriate amount of time, that's OK too.

The point is that we get to decide, not a mechanistic process, in the case
that someone knows of a good reason to keep that code around for a few
releases more.

-- 
John Wiegley                  GPG fingerprint = 4710 CF98 AF9B 327B B80F
http://newartisans.com                          60E1 46C4 BD1A 7AC1 4BA2