[_fbsd_chat_] why the different eol lengths ? [ was: Re: [FreeBSD-Announce] HEADS UP: FreeBSD 7.2 EoL coming soon ]

[Dag-Erling Smørgrav]

spellberg_robert <emailrob at emailrob.com> writes:
> if version "x dot y_plus_one" is "better" [ for some definition of
> that word ] than version "x dot y", then why is support for the
> "better" version being terminated, by --design--, before support for
> the "less_than_better" version ?

It is a simple matter of resources.  We have limited manpower and can't
afford to support too many releases at once.  Even the current system
(one year for normal releases, two for extended releases) strains our
resources; we currently support five releases (6.4, 7.1, 7.2, 7.3 and
8.0).  Support for 7.2 ends on June 30th, but 8.1 is right around the
corner.

Note that this is how "everybody" does it (for suitable values of
"everybody").  Ubuntu, for instance, supports regular releases for 18
months and "LTS" (long-term support) releases for 36 months.  Every
fourth release (the spring release in even-numbered years) is an LTS
release.  And they're not even all-volunteer - if they were, they'd
probably cut down to 13 / 25 instead of 19 / 37.

> in the sequence "long_date_a, short_date_b, long_date_c": if i am
> running "a", then why should i upgrade to "b"; if i am going to have
> to upgrade to "c", anyway ?  i may_as_well wait for the release of
> "c", in the first_place.

Large (i.e. corporate) users usually won't consider upgrading to a
particular release until it's been out for six months or a year and
they've had time to test it thoroughly.  If we didn't support at least
some releases for significantly more than a year, they'd abandon FreeBSD
entirely.

I have customers who are still running Ubuntu 8.04 LTS and have barely
started evaluating 10.04 LTS.  They probably won't start deploying it to
production systems until late this year or early next year.

(sadly, I do not currently have any customers running FreeBSD)

DES
-- 
Dag-Erling Smørgrav - des at des.no