[Qgis-developer] Stress about release plans

[Zoltan Szecsei]

On 2014/07/22 17:45, Paolo Cavallini wrote:
> Il 22/07/2014 17:17, Bo Victor Thomsen ha scritto:
>
>> My suggestion is that*one* of the three version cycles is replaced with the following:
> Sorry Bo if I appear rude, but good ideas are not what we are missing.
> Manpower (~=money) is what is missing to implement them.
> Therefore: funding welcome, be sure we can devise a suitable strategy to solve these
> problems once we have suitable resources.
> All the best, and thanks for your thoughts.
>

Wow, I could be so off the mark with this, as I am not even aware if 
there is a regular donor for QGIS, or just ad-hoc income - but perhaps 
me coming from the outside, might give you fresh ideas to mull over.

So,

The guys who have been long term dev contributors to QGIS (effort, not 
money) should surely have some idea what effort it might take maintain 
an LTS version.
Perhaps they could agree that for the 1st LTS version, one or two of 
them would take on the responsibility for maintaining that LTS version 
(yes, they might need the guidance of the people who wrote the part of 
code "that blew up").
For the next LTS version, this responsibility could move to other devs, 
and so the burden of LTS maintenance could rotate amongst suitably 
qualified and willing devs.
Then, given that QGIS is a mature project that is widely used, is there 
not some organisation that would be willing to fund the LTS maintainers 
on a monthly retention basis? (a retainer, rather than a "fee per fix" 
will offer small but regular income for the dev.)
Canonical?
Would OSGEO have some contacts?
Maybe set up a special corner of the QGIS Web page where you could show 
the Company logo of anyone prepared to sponsor a particular LTS cycle? 
(This last thought would mean that QGIS would have to fund-raise for 
each LTS cycle - not great :-(  )

If you coincided the LTS versions with (say) Ubuntu, that would also be 
useful for stability, but that could mean that the LTS maintenance Devs 
would have to deal with backports so that QGIS does not become too old 
before the next LTS cycle.

Note also that most new features come with plugins, so as long as a new 
plugin is also certified against the prevailing LTS version, QGIS users 
won't have to suffer with a (too much) out of date QGIS.

Anyway, I hope my thoughts will stimulate ideas on how to handle this, 
but in short perhaps decide in this order:

 1. How long between LTS cycles (maybe pick 2 or 3 scenarios)
 2. How many devs need to be available for ad-hoc bug-fixes for a
    particular LTS cycle
 3. How these (willing) devs could be cycled over different LTS periods
    so they don't get bored and stuck with maintaining old code.
 4. What effort (read money) would they want _as a retainer_ (rather
    than "fee per fix")
 5. How to get the money for the retainers

Regards,
Zoltan


-- 

===========================================
Zoltan Szecsei PrGISc [PGP0031]
Geograph (Pty) Ltd.
GIS and Photogrammetric Services

P.O. Box 7, Muizenberg 7950, South Africa.

Mobile: +27-83-6004028
Fax:    +27-86-6115323     www.geograph.co.za
===========================================

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/qgis-developer/attachments/20140722/4455486c/attachment.html>