Re: Some comments about the paper
|--==> "MA" == Miguel A <Arévalo <email@example.com>> writes:
Miguel, thanks for having replied me on list, actually my post was
intended for the list, but just missed to keep it in Cc.
MA> El mar, 28-09-2004 a las 15:06 +0200, Free Ekanayaka escribi~:
>>|--==> "MA" == Miguel A <Ar~valo <firstname.lastname@example.org>> writes:
MA> at T0 + 1 year:
MA> freeze := stable
>>Perheps you mean free := testing ?
MA> Yep, sorry.
MA> Time based release cicle is good (this is of course my opinion) but
MA> don't flame me on these, the scheme can be perfectly T0 + desired
MA> features ready.
>>Time based is good for me too, because with feature based you
>>typically miss releasing because of new feature requests. BTW GNOME
>>has time based releasis (at least AFAIR): they decide a date and who's
>>in is in and who's out is out.
MA> Yep, I think this could be very good for Debian, so at the moment
MA> testing becomes freeze if something is not ready, then it will be left
MA> for the next iteration (but remains in testing).
I forgot to mention another interesting option:
DD -> unstable -> testing -> releasable
where releasable it's testing with only RC-free packages.
at some moment we either "freeze" releasable or copy it in a separate
freeze suite (freeze := releasable), which would behave as you
This way a reasonably stable and up to date set of package would
always be available, letting a CDD to schedule independently its own
freezes and releases.