Developer Meeting Notes: October 12, 2014
Ton Roosendaal shares the notes from today’s meeting in irc.freenode.net, #blendercoders.
1) Will there be a 2.72a update?
- There have been a couple of fixes that deserve to be released, just to make it rock solid.
- The team will check on what to merge in the next days, wednesday we could call for a build.
2) Targets for next release
- The planning proposal.
- Defining the final release targets next week might be too soon, several developers prefer more time. This then will be two more weeks, because of Blender Conference.
- Please don’t forget to help Martin Felke!
- Antony Ryakiotakis and Jason Wilkins will review if it’s possible to apply the Viewport project for 2.73 still.
- Request for review, adaptive sampling for Cycles.
3) other projects
- Lukas Toenne’s hair simulation and editing work might make it for the next release.
- Sergey Sharybin is recoding Texture Nodes, simpler and thread safe.
- Buildbot was exposing binary testbuild (branch) downloads from unclear origins, these get hidden now.
Will 2.72a include proper sm5_2 support for the fancy new GTX900 cards?
no, in the latest builds or in 2.73 -> see here : http://wiki.blender.org/index.php/Dev:Ref/Release_Notes/2.73/Cycles
“GPU Rendering
Geforce 9xx GPUs are now supported.”
I know all about the new kernels, currently using experimental builds. The issue is that hardware forward compatibility is considered a “feature” rather than a “bug”, and therefore never considered for a/b releases despite the bug being very serious for CUDA users with new cards.
Oh please please please keep going with that adaptive sampling! Right now I have to do it manually by hand. I render scenes where usually 90% of the scene is getting fairly direct light, but the remaining areas getting indirect light come out noisey unless I increase the sample rate dramatically. So I usually have to render small sections of the images with higher quality and render the whole image with a lower quality, just to speed up a render. Where I work I don’t have time to wait 2 hours for a render, if it has to be done in 15 minutes, it has to be done, even if it doesn’t look perfect. So this would be just heavenly for me!
I have just downloaded a test build and I am very encouraged by initial test results. I may just call this my render build until it makes it into trunk
Mind if I ask for a link to the test builds? I’d love to give it a go.
I’m using the Win 64 build from the developer found on post #47 here:
http://blenderartists.org/forum/showthread.php?350771-Cycles-Adaptive-Sampling/
Be sure to copy the missing DLLs from a standard Blender 2.72 distro mentioned on post #79 in the same thread
How about handicap accessibility be a target for the next version? Streamline workflow so that a person with only one arm can use your program with a minimum of frustration. It can be done in Maya with shelves and pie pie menus. How can someone do it with Blender? If Ton Roosendaal got into a car accident tomorrow and lost his arm, would he be able to productively use Blender in the next revision?
I may not be able to relate, but I think I understand the feeling. With all the hot keys required to use the program, sometimes two arms doesn’t feel like enough, I can’t even imagine what it’s like for someone with only one :(
I think more there needs to be more proposals made to address this.
Handicap accessibility is a good but giant step. It’s a pretty extensive process of research, development, and testing.
It’d probably require a fundraiser, just to assemble a team of include designers and researchers who specialize in this line of work–it’s not an easy or quick thing to tackle effectively with a group of variously-gathered developers with no experience in the field of accessibility design.
I have a bit of familiarity with researching the world of handicap accessibility with game design, enough to know that when you’re doing something like it, you’re taking an extensive commitment, in terms of research and development.
I support the request for better accessibility.
Let’s not start an “argue” about “Blender UI” like there was that kind of discussion before, but let’s keep in mind there are artists with impairments.
It would be nice if this is considered as important as the necessary (and very useful!) improvements we see with every new release.
Of course it takes time and will be a process, but I’m pretty sure it’ll turn out beneficial.
Why not make it a goal on the Blender Roadmap?
What you want the team to do serious UI development, I’m sure they will once pigs start to fly. Sadly like the majority of open source projects UI is of zero importance to them.
sure hope felke gets the help he needs to get this in trunk.and thanks for stickin with it this long martin, im sure its tons of work, good luck with review, and hopefully inclusion in 2.73.mabye i hope.
what about movin origin with one klick …i mean a behaviour like other element too(see bellow)…and scale just from one side like in the old truespace? The User Open3Dart developed a nice tool under http://www.adrive.com/public/MjwnNg/PivotMode.py , but it would be nice to see an optimized and native implementation in blender. thanx
do you know the shotcuts for pivot points “cnrl period , ctrl comma , period , comma , and
alt comma , alt period. not really that difficult once you get used to it. not trying to argue, but it is pretty simple.
any news on the 2.72a release, when will it actually be out ? anyone ? just asking. :)
2.72a is out -> http://download.blender.org/release/Blender2.72/
got it ,thanks.little impatient i guess.