[Qgis-developer] Help us plan for QGIS 3.0

[Neumann, Andreas]

Hi, 

QGIS is now used by a lot of professional users worldwide. A lot of them
have contracts with QGIS developes/companies and expect that they can be
delivered within a few months. If we move now to QGIS 3.0 on such short
notice, without any concrete plan/warning, means such projects can only
be delivered in early 2017, which might be a problem for many, including
me and my employer. We need a few improvements in QGIS forms and widgets
in the 2.x branch. We can't wait for them to be delivered only in 2017.
CAD import would also be nice to have soon - though that is not equally
important compared to the forms improvements. 

I have to repeat again that Anita (also board member) promised that the
board will inform all involved parties (e.g. devs, professional users)
well in advance with such decisions. Now changing immediately to QGIS
3.0 on such short notice without a clear plan is exactly the opposite of
good communication of the QGIS board with the user and developer base. 

Because of this, I would like to have another 2x release, like Matthias,
Hugo and Nyall suggest. We can then, as Mathias suggests already
introduce Python 3 support in parallel with Python 2 support, which is
an important ingredient of the upcoming QGIS 3.0 version. It would then
still give us 8 months to mature the 3.0x version before the next LT
release. 

Hugos suggestion to have 2.16 as the LT release instead of 2.14 would
also be fine and acceptable for me. Moving to 3.0 immediately would not
work well for us. 

Thanks, 

Andreas 

On 2016-01-19 11:08, Paolo Cavallini wrote: 

> Il 19/01/2016 10:38, Richard Duivenvoorde ha scritto: On 17-01-16 23:03, Tim Sutton wrote: [1] http://blog.qgis.org/2016/01/17/help-us-to-plan-for-qgis-3-0/ 
> Dev's,
> 
> Nobody else?

My suggestion:
* release 2.14 LTR
* move immediately to 3.0
* backport all fixes to 2.14, as usual for LTR
* allow for exception for new functions, on a case by case basis, if
these are especially important, of limited impact, and needed soon, well
before the expected release of 3.0 (I know this is against the LTR
concept, that's why I think it should be exceptional)
* provide help for plugin authors to migrate them to 3.0 API and libs.
Given our limited resources, and the need to move forward to follow the
availability of new libs on some platforms, I think the sooner we start
the better.
All the best. 

  
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-developer/attachments/20160119/95895fe3/attachment.html>