Re: post-release package update policy
>>>>> "Matthew" == Matthew Swift <firstname.lastname@example.org> writes:
Matthew> release" last week, as some have wished. The last couple of
Matthew> days of debian-changes tell me that versions of two new
Matthew> packages are available now whose only modification was to
Matthew> include new or improved documentation. Another update for a
Matthew> network package included significant security improvements.
Matthew> I think anyone downloading a release today -- including
Matthew> CD-ROM vendors and new users -- would want a Debian with
Matthew> those new packages, not last week's hypothetical "release,"
Matthew> however official.
All official means to me is that an effort has been made to document
what is there. Functional stability is nice, but is, IMHO, very much
secondary. Instead of having the debian coordinators agonize
over some arbitrary determination of `stability', why not just
stop when you say you are going to and then post to debian-users
retrospectively as problems come up.
I'm all for the notion of an evolving system, and I think that all
users should become immediately aware of this aspect of Free software --
people get too whiney and parasitic, otherwise. With a broad but
shallow set of maintainers, it seems especially important to encourage
people to take the initiative. It seems to me that delaying serves
all the wrong people.
The absolute worst thing would be to have a setup where it wasn't
clear to the naive user what updates had been made to their R6. Even
worse would be slightly different versions on different CDs.