[Bf-committers] Proposal for 2.36 release

[Juan Julio Peña Mena]

I believe that shorter Release Cycles not only are not the solution, but
also will work againts Blender's public image.  

Theeth once said in a meeting (in joke tone), that the Blender Forum at
CGTalk is somehow the Blender Embassy for the CG-Country, since it is
the closest that we have to a neutral forum where we get to see the
point-of-view from by-passers who are not that passionate about our
community, nor that biased about Blender itself.   What I'm about to
write comes from my observations on "The Embassy".
--------------------------------------------------
As things stand now, every time a new version is released it gathers a
lot of public attention, since it is something that happends few times
in a year, and everytime it happens it comes with amazing new features -
not to mention that it make it to the cgnetwork news headlines.  
Sometimes (as with 2.34) you can witness the expectation really growing
stronger even from a month before the actual release.  That kind of
expectation is really paying off to attract new blenderheads..  Since
2.30 till now, the Embassy has passed from been a neart-death-comatose
forum to a quite lively place, now averaging 40 something active threads
at any given time. (a thread is considered active if it last post is
less than 30 days old).

Now take Wings3D for comparison, they realese new versions quite ofthen,
and people barely notices or cares about the release, since it is
something so common.  Even on their own forums the new releases are
usually ignored.
-----------------------------------------------

I agree that Release candidates should be more "officially" sponsored on
the main community sites.  I, one of the so-called "official beta
testers" myself, often choose to ignore some test builts posted on the
forum because they are ... how to put it in words... fan-produced.  If
they all where produced by the same developer that for sure will compile
the final released version the apreciation would be a lot different. 
But as experience tells me, many bugs comes from different
compilation-time scenarios that will vary from developer to developer,
so when you have very limited free time you really wonder "is this
particular test built really worth the testing time and effort?".  I
guess that many of the t.b. forum visitors think alike. 

That lead me to the Regression Test Suite... anything that could be done
to make the suite faster to perform would be most wellcome.  Maybe a
fully automatic test is just too much, but if anything could be done to
expedite the suite I am 100% with it.

Also, I agree that the testing step on the cycle should be prolonged,
not all proyects need all features.. and I preffer to test Blender with
real work whenever it is possible, but that takes more time.  About
issuing a "tester registration form".. honestly I'm neutral about that,
but if it means that we get more "oficial" testing subjects then count
with my support for it.

Apollux.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://projects.blender.org/pipermail/bf-committers/attachments/20041123/0d550f2f/attachment.html