Just so I am clear is providing LTS part of this proposal to change the life span of the stable branch? Or something that is going to be considered separately?
···
On Fri, Jan 24, 2014 at 9:18 AM, Chris Holmes <chomie@anonymised.com> wrote:
All that makes good sense. I wasn’t actually advocating for putting more resources in to maintaining really long releases, and indeed I think the 1 year cycle is good. Was just suggesting the notion of not doing LTS for every release. So what I was aiming at was less work for developers. Basically instead of doing the LTS for every release, of old bug fixes, we’d just do it for say evens. So then when people are like ‘why is this simple bug fix not in 2.3.x’ you can say that only 2.4.x is LTS, so it’s the only one that gets the old backports.
And then from there just promote the idea of LTS a bit more for stable people. Every year they’d upgrade to the new LTS, when the next even gets to LTS status, instead of every six months. So like Jonathan is saying, we wouldn’t have LTS and optional LTS, there’d just be one LTS that stretches out.
CenturyLink Cloud: The Leader in Enterprise Cloud Services.
Learn Why More Businesses Are Choosing CenturyLink Cloud For
Critical Workloads, Development Environments & Everything In Between.
Get a Quote or Start a Free Trial Today.
http://pubads.g.doubleclick.net/gampad/clk?id=119420431&iu=/4140/ostg.clktrk
Geoserver-devel mailing list
Geoserver-devel@anonymised.comsts.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/geoserver-devel
–
Justin Deoliveira
Vice President, Engineering | Boundless
jdeolive@anonymised.com
@j_deolive
On Fri, Jan 24, 2014 at 7:02 AM, Andrea Aime <andrea.aime@anonymised.com> wrote:
On Fri, Jan 24, 2014 at 5:16 AM, Chris Holmes <cholmes@anonymised.com> wrote:
(apologies for long silence on stuff, have been taking some time away)
I think one thing we might consider is only doing LTS releases on like evens or odds, to potentially ease things on us more. The one thing that strikes me on these schedules is that LTS is just kind of ‘old’. The Ubuntu LTS concept is that every 2 years they release one that they commit to for much longer (5 years). So the conservative people just do that, and know they’ll always get fixes. See https://wiki.ubuntu.com/LTS
Yep, familiar with the concept (using a Ubuntu LTS myself).
However, not sure this can apply to GeoServer for a couple of reasons.
The main one is that we’d be trying to bite more than we can chew, in terms of resourcing.
Let’s imagine we had a 1.5 years old LTS right now, that would be the 2.2.x series, and generally speaking, something 4 branches behind the current development series.
The development of GeoServer is still going at a rather high pace, doing a backport of a fix becomes more and more hard as we just back in the branches,
while doing a cherry-pick from dev to stable (one branch jump) normally works without significant conflicts, a 4 branches jump will result in conflicts very often,
and in a need of full rewrite of the patch also often enough.
What does that mean? That the bugfixes in questions are realistically going to be backported only if you are under contract to do so, so, few of them.
Which brings me to the second source of concern: who is going to be interested in a 1-2 years old release?
In my experience, no one, and I’m not talking hyphotetically, GeoSolutions tried with “GeoServer enterprise”, and it basically got no traction:
all of our customers want something out of the recent series, and then they want that to be maintained.
A 1-2 years old releases is just lacking too much, mostly because we are still going at a high pace.
That’s why we are pushing a 1 year lifetime instead, it gets long enough for conservative people to use, but not so long that people will just
ignore it due to lack of too many new features.
Point in case, the new precise GetFeatureInfo… there is a lot of interest for it, and we’ve been asked already a few times “why is it not in 2.4.x?”
(the answer is, because it’s not a simple bug fix, it’s a major refactor).
Something like this lowers the interest in older releases a lot… a few items like this and the interest in a 1 year old release, 2 cycles
of new features behind the curve, is already rather small.
Imho, LTS are great, but for stuff that is mostly rather stable and not fast evolving as GeoServer still is, 2 years cycles are just impractical
from both the developer point of view (difficulty of backports) and from the audience interest point of view (which also means, difficulty
in gathering funds to justify the extra effort to keep a LTS up this long).
Cheers
Andrea
–
== Our support, Your Success! Visit http://opensdi.geo-solutions.it for more information ==
Ing. Andrea Aime
@geowolf
Technical Lead
GeoSolutions S.A.S.
Via Poggio alle Viti 1187
55054 Massarosa (LU)
Italy
phone: +39 0584 962313
fax: +39 0584 1660272
mob: +39 339 8844549
http://www.geo-solutions.it
http://twitter.com/geosolutions_it