The major OS competitor has moved to a 6 month release cadence, so that needs to be taken into account.

I suspect the easiest way to have a longer supported system is by NOT having a longer supported system. For the case of desktop hardware vendors something like a constantly updating silverblue might fit better.

It could be a continuously updating system "automatically" moving to the next release shortly after the next fedora version is released. As it is image based breakage should be harder to occur.

It will be no more or less faster updated than competitor operating systems that manufacturers use.

On Thu, 15 Nov 2018, 00:15 Ben Rosser <rosser.bjr@gmail.com wrote:
On Wed, Nov 14, 2018 at 6:04 PM Stephen John Smoogen <smooge@gmail.com> wrote:
>
> On Wed, 14 Nov 2018 at 16:03, Ben Rosser <rosser.bjr@gmail.com> wrote:
> >
> > On Wed, Nov 14, 2018 at 2:55 PM Stephen John Smoogen <smooge@gmail.com> wrote:
> > > From what I have talked with in the past.. 3 years is their bare
> > > minimum and 7 is their what we really want. It usually takes the
> > > vendor about 3-6 months of work to make sure the OS works on their
> > > hardware without major problems and then they want people to buy
> > > support contracts for 3-5 years where the number of problems needed in
> > > year 3-5 are none. [This means that they want to have Fedora N for 3-6
> > > months before their laptops ship with it. So you ship them a frozen
> > > preload before you release to public. They also want any shipped to
> > > 'last' for the warranty cycle because trying to deal with update
> > > questions when N eol's in the middle costs them a lot.]
> >
> > If 7 years is what manufacturers really want, then it sounds like
>
> Well they also want a Ferrari and all support to be done upstream for
> free. 7 is usually their counter to 13 months. You start going down
> there to find that what they really settle for will be 3-4 years as
> most people don't extend warranties that long.

Well, even so, 3-4 years would be a pretty long time.

My point about EPEL was that, Fedora currently does produce a
long-term-support type product (admittedly for another distro). It's
EPEL.

Except we don't really push it. EPEL is an opt-in thing, which means
lots of packages don't have EPEL branches-- possibly because the
maintainers didn't want to commit to maintaining a package in a
long-term-support type environment, or possibly because the
maintainers never thought or bothered to create an EPEL branch, or
possibly because there are too many dependencies that don't exist in
EPEL and tracking down those maintainers isn't worth the time and
effort. I know that I would package more things for EPEL if I could
reuse Fedora specs with a minimum of fuss and I didn't have to spend
time getting a bunch of dependent packages built.

It is not clear to me that Fedora having two long-term-support type
products would be a good idea, as I am not sure that we have the
resources to maintain *one* at the moment. That makes me think we'd
want to tie a hypothetical "Fedora LTS" directly to RHEL/CentOS/EPEL
somehow, and find a way to reuse the work for both efforts.

Ben Rosser
_______________________________________________
devel mailing list -- devel@lists.fedoraproject.org
To unsubscribe send an email to devel-leave@lists.fedoraproject.org
Fedora Code of Conduct: https://getfedora.org/code-of-conduct.html
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: https://lists.fedoraproject.org/archives/list/devel@lists.fedoraproject.org