Quoting Jeremiah C. Foster (2019-03-14 21:29:01)
On Thu, 2019-03-14 at 15:10 -0500, Omar wrote:
On 3/14/19 1:52 PM, Jonas Smedegaard wrote:
Quoting Jeremiah C. Foster (2019-03-14 18:57:44)
What do those folks on this mailing list think? Should we keep PureOS Green on Debian (Buster) Stable?
Above is strongly tied the related question of what to do about cravings for exciting new $stuff as Buster (non-)evolves to become steadily more boring over its multi-year lifespan.
I might give you another perspective from an intermediate user. What some of you 'OS nerds' ;) consider boring, I'm guessing the majority of our customers see it as a very functional, cool as-is tool to get things done. As long as privacy and security improvements don't get stagnant... And any customer that may be as advanced as you guys, will know the ways to make it un-boring :)
+1
I think the majority of our enterprise user base will feel the exact same way.
Ok, so the majority of users would be happy today with...
* Linux 4.9 * GNOME 3.22 * LibreOffice 5.2
Great!
I am not grumpy about that at all. I thought I heard others in Purism - people _not_ in PureOS team - be grumpy about especially GNOME not being new and shiny - but possibly I misunderstood or our real users are far different and easier to satisfy. Just great :-)
Do we...
a) Tell users to wait for it to become boring enough?
Yes.
Great! I am happy to be mistaken.
b) Maintain a local fork as .deb in PureOS for each wish? c) Maintain a local flatpak for each wish?
flatpak is going to be installed on the system so those who want cutting edge can turn to that.
Do you confirm that we _will_ hire enough people to handle locally maintained flatpaks?
d) Tell users to include .deb/flatpack maintained elsewhere?
...or did you mean that we rely on externally maintained flatpaks?
With a) I say yes let's do it. But I expect others in the company to not really want that option for several years, not even for enterprise users.
What evidence supports this?
Evidence?!?
I have absolutely zero evidence about _any_ of the 4 scenarios: PureOS never explored any of them, so this is all pure speculation.
I've generally received positive feedback when talking about stability with the move to Debian Stable from other parts of PureOS. I have received some anecdotal evidence along the lines you've stated, that 5 years is too long, but two years may be reasonable. H
So we bet on Bullseye being stable in two years, or we switch PureOS to being a rolling release in two years, or why mention that timespan?
I cannot promise a certain timeframe, but have evidence of historical pace of Debian releases: https://timeline.debian.net/
What about providing a dist-upgrade to Bullseye when it is stable?
What about it?
As with everything else, there are the same 4 scenarios.
For scenario a) Debian supports dist-upgrade from one stable to next stable, so as long as PureOS tracks stable releases of Debian then PureOS is equally dist-upgradeable.
Any parts deviating from Debian is our headache to ensure upgradeability for, however. Currently we have practically zero testing - manual or automated - because we are currently a rolling release where upgrading is eternal: Anyone had problems last week? Just try again this week, perhaps the problem fixed itself...
But if I understand correctly, we eliminate maintenance issues, including upgradablitity, by using flatpak for anything where Debian has evidence of maintenance being required.
- Jonas