Hi Marco -
First of all, I would like to thank you and the PSC for keeping the
concept of LTR. I was worried after the discussion on developer list
recently.
My personal opinion - based on empirical observations of QGIS usage and
as a member of the steering committee of the Danish QGIS user group - is
that a vast majority of QGIS user (at least in Denmark) is using a
recent version of QGIS LTR as their primary QGIS system. And the
adventurous users or users in need have a parallel installation of a
later, non-LTR QGIS on the same pc.
It would be detrimental to the effort of expanding or even keeping the
QGIS user base if you abandoned the LTR concept.
Regarding your questions:
* The reduction of number of 12 yearly point releases to 4 has no
practical disadvantages and will reduce the work effort of getting
the releases out of the door. I don't know a single organisation
that install all the point releases. 4 is fine.
* The saved time could - and probably will - be used for more
extensive testing the 4 releases. This will result in overall better
releases with fewer errors than before.
* And - again - my experience is, that the vast, vast majority of QGIS
users is on Windows. It makes sense to maximize the pre-release
testing effort on this platform.
* AFAIK, my installed QGIS is communicating with QGIS.org using the
newsfeed facility. You can turn it off if you want to. I don't see
anything problematic in delivering information of the operating
system type together with the QGIS version.
So a big +1 and thumbs-up on all 4 proposals.
Med venlig hilsen / Kind regards
Bo Victor Thomsen
Den 20-11-2021 kl. 08:00 skrev Marco Bernasocchi:
Dear community,
since their introduction, QGIS LTR versions have gained wide-spread
popularity [0]. LTR point releases have always followed the regular
monthly release cycle [1] creating a constant flow of point releases
that likely exceeds the update frequency that LTR users desire and
actually roll-out.
To further improve our LTR offering and ensure the reliability of our
installers, we propose to introduce pre-release LTR package tests.
These semi-automated and manual tests will be performed during a
two-week period between packaging and announcement of new packages.
The 2022 budget proposal will account for the additional costs of
pre-release LTR package tests. To stay within a reasonable budget and
allow for an appropriate number of tests, we propose to reduce the LTR
point release frequency to quarterly (four times a year) and initially
limit testing to the platform with our largest user base (Windows).
If accepted, the new quarterly schedule would start as soon as the
whole testing infrastructure is ready. This would probably be
introduced stepwise during the 3.22 lifecycle and become definitive in
the 3.28 cycle.
Together with this proposal, to quantify the userbases by operating
system, and be able to make data driven decisions in the future, the
QGIS network user agent now includes a generic string mentioning the
user’s operating system. You can read more about the change here [2].
We are looking forward to hearing what you think about these news
plans and your ideas how they may be improved.
On behalf of the PSC,
Have a great weekend
Marco
[0] https://blog.qgis.org/2020/04/02/ltr-usage-survey/
[1]
https://qgis.org/en/site/getinvolved/development/roadmap.html#release-schedule
[2] https://github.com/qgis/QGIS/pull/46090
--
Marco Bernasocchi
QGIS.org Chair
OPENGIS.ch CEO
http://berna.io
_______________________________________________
Qgis-user mailing list
Qgis-user@lists.osgeo.org
List info:https://lists.osgeo.org/mailman/listinfo/qgis-user
Unsubscribe:https://lists.osgeo.org/mailman/listinfo/qgis-user
_______________________________________________
Qgis-user mailing list
Qgis-user@lists.osgeo.org
List info: https://lists.osgeo.org/mailman/listinfo/qgis-user
Unsubscribe: https://lists.osgeo.org/mailman/listinfo/qgis-user
