Good luck everyone, this is promising.
The reason why Blender doesn’t have true asset management yet, is because of all of the steps required beforehand so Blender can even have one (a few of which have already been completed and are in master).
Considering what Bastien had to start from code-wise, it’s a much larger project than you think and a student would not have been able to do it during the GSoC period as mentioned already.
It’s data storage optimizations, not texturing in general. So no, no procedural stuff. There’s other things that can use the planned optimizations besides image textures though, like smoke/point density grids, UVs, etc. Proposal writeup is here: https://wiki.blender.org/index.php/User:DingTo/GSoC_2016/Proposal
but tl:dr (in order of schedule):
-No more texture count limits on newer Nvidia GPUs
-Grayscale textures like bump/roughness maps, masks, or voxel datasets can be stored as-is instead of converted to RGB, cutting their memory use to 1/3rd
-Add a half-float data type for 16bit image textures and general storage data that doesn’t need the precision of 32bit (UVs and vertex normals are the given examples).
-Improve logging and memory stats so Cycles can give more useful advice in optimizing your scenes.
-Support the OpenImageIO image cache in SVM CPU mode. Not 100% sure if this includes tiled/lazy loading initially. GPU support is more complicated (requires rolling our own tilecache) so don’t expect it during this project.
I am aware that it is not simple, because blender did not have this from the start. But when that is said, id there are so many steps onvolved, why can’t they take it from Microvellums Fluid designer that allready have all this implemented? it open source also so they are allowed to take the code from there.
Pretty sure it is 1/4… as it stores RGBA not just RGB
Not sure why you want to discuss that in this unrelated thread. However, do you think the Blender developers do not consider different approaches before they start to implement something?
I look forward to seeing improvemnts to the camera tracking part of blender. Cant wait to see
- tianwei37 work on this. God luck to all of you. Meck
Layer manager is the best project for me 
finally .
What are those steps?
The first one was rewriting the file browser window (which was merged a while ago), I think there was also another step (that I don’t remember) that is also in master.
Here are the current ongoing projects related to asset management
https://developer.blender.org/diffusion/B/repository/asset-engine/
https://developer.blender.org/diffusion/B/browse/asset-experiments/
These are not small projects by any means, so again, it shouldn’t be any surprised that it’s not a GSoC project.
Another big sub-project of the asset manager was/is the ID remap, https://developer.blender.org/diffusion/B/browse/id-remap/
I’m looking forward the Sculpting improvement!
Is there a link page on the blender wiki to all of the GSoC projects? I know that each student has a proposal an update page, but they seem to be uncollected.
Cool! Thanks XRG, I couldn’t find this anywhere.
Any news with Mantaflow GSoC 2016? Can we see build for Windows? Will Sebastián Barschkis continue this project?
Best,
Chris
The final presentation of the GSoC projects on code.blender.org
https://code.blender.org/2016/08/summer-of-code-2016-results/
8 out of the 10 projects were successful it looks like (hopefully the BF can encourage the students to stick around so as to get everything into master).
Almost all projects completed or nearly completed, this is great. It is the best GSOC I have seen for now. I hope that all projects can join the trunck. I can not wait to test the new layers 
Congratulations