Hi All,
We've been discussing the next stable release, OpenSceneGraph-3.0, on
and off for a year now, and it's been coming soon for rather too long.
I believe it's time now to make a push to get 3.0 out the door, and
if we get the testing done, code refined and not attempt any ambitious
new feature development we should be able to get it out in the next
three weeks - in time for a well deserved rest at the end of year ;-)
Getting 3.0 before Christmas is a very competitive goal. To achieve
this I will need to be pretty strict about what new codes and fixes go
in and what features we attempt to finish. OpenGL ES support is well
established, iOS support is well on it's way to being ready, and the
new serializers work very well in most usage models. All of these
should be possible to tidy up in time for 3.0 without undue risk to
the competitive schedule.
One feature I have long assigned as a MUST HAVE for 3.0 was shader
composition, and I and others have done work in this direction, but as
things stand no one implementation is flexible and complete enough to
say that it achieves the goals I wanted for it. I have been busy with
too much other work in the last three months to complete what I
started with what is now integrated with svn/trunk, what is there is
about half done. For limited usage models it can be used
successfully, and for more traditional usage of the OSG it doesn't
effect usage so while it's incomplete I don't believe it's presence in
svn/trunk need hold back 3.0. I have to be realistic though, there
isn't anyway I or anyone else can perfect such a cutting edge feature
in the next few weeks, so rather than try to chase the goal of having
full feature shader composition, I feel it we just need to let this
feature take a back seat till after 3.0 and shift it to a goal for the
follow up stable release 3.2.
In prep for 3.0 I will start pushing out one or two more dev releases,
and then go a feature freeze and the 3.0 branch, from this a series a
release candidates will be made, till the final one goes out the door.
A provisional schedule will be something like:
Friday 3rd December OpenSceneGraph-2.9.10 dev release
Friday 10th December OpenSceneGraph-2.9.11 dev release
Tuesday 14th December OpenSceneGraph-3.0 branch made
Wednesday 15th December OpenSceneGraph-3.0-rc1
Friday 15th December OpenSceneGraph-3.0-rc2 (if required)
Monday 19th December OpenSceneGraph-3.0-rc3 (if required)
Wednesday 22nd December OpenSceneGraph-3.0-rc4 (if required)
Somewhere between 19th to 24th December official OpenSceneGraph-3.0 release
Hopefully this competitive schedule won't sound too crazy. Lots of
testing will be required from myself and others in the community - a
busy time for us all even without adding extra release work. The more
testing we can do of the OSG and how well it works against our
applications the more we'll be able to fix problems and refine it for
the final stable release.
When working within this time frame we all need to think from the
perspective of minimizing the risk on the schedule and prioritizing
code robustness over feature development, so it's no time to start
proposing wouldn't it be great to have feature x for 3.0. It's about
getting 3.0 out the door, further new features will be for 3.2 and
beyond.
This is just a first post in a series for working towards getting 3.0
out the door. If you have thoughts/opinions/suggestions about getting
3.0 smoothly then please post them here.
Thanks in advance for you help,
Robert.
___
osg-users mailing list
osg-users@lists.openscenegraph.org
http://lists.openscenegraph.org/listinfo.cgi/osg-users-openscenegraph.org
