super-drafty F28 and F29 schedules

[Matthew Miller]

It's probably no surprise if you've ever heard me talk that I really
think hitting early May / late October is important for our release
cadence. Those specific dates aren't magic, but they avoid some big
public holidays, and the key thing is that if we're consistent with
landmarks, it makes planning easier.

I took a look at the planned F27 schedule
https://fedoraproject.org/wiki/Releases/27/Schedule and sketched out
what the same time periods would look like with a May target for F28,
and then repeated again for F29, making very drafty preliminary
schedules:

https://fedoraproject.org/wiki/Releases/28/Schedule
https://fedoraproject.org/wiki/Releases/29/Schedule

and, really, I think it looks very nice. There's a built-in week for
both Beta and Final to slip if need be, and, realistically, if we end
up going a week or two past the planned GA, we're still in decent
shape in May and November.

The key thing, though, is that this is predicated on "No More Alphas"
actually working. That in turn has two big parts.

First, there is gating from rel-eng and QA in progress here: 
https://fedoraproject.org/wiki/Changes/NoMoreAlpha (Note that this is
compose/validation gating, not the CI stuff we're also talking about
separately.) That's key in keeping the release basically stable. But
there's another part:

Unless we want to have very soft release criteria and ship stuff we
know to be broken (or suspect but haven't completely checked out), we
need changes after we branch (and especially after the beta is
released) to follow the stable releases updates policy --
https://fedoraproject.org/wiki/Updates_Policy#Stable_Releases. That
way, we have a fighting chance of identifying blockers with enough time
to get resources directed to fix them without indefinite delay.

Hopefully, by the time we are at F28, Modularity will provide a way for
us to offer faster streams for people who want them -- but let's also
focus on stable releases. The ironic effect of this is that we can get
new software to people *faster*, because we'll be able to more reliably
hit the six-month ship dates, rather than taking eight or nine months.

-- 
Matthew Miller
<mattdm(a)fedoraproject.org&gt;
Fedora Project Leader