Given that the development of the actual Collada format is “dead” for all practical purposes, I’d expect that “maintenance” of the library would more or less be 1) fix any serious bugs or security issues uncovered in the library itself, 2) update 3rd party deps that the library itself embeds/uses, 3) make it build with current compilers / C++ standards.
If that is needed, I can “volunteer” to do the maintenance. With the usual caveats, like I already managed to introduce a collada bug in 4.2.0 while cleaning up 3rd party dependencies (fixed in 4.2.1, but required platform maintainers to rebuild the lib). It would actually be easier if the library itself was vendored under extern directly within blender.
not thrilled about extern existing in the first place. making every single dev build some potentially expensive library (that rarely changes) every time they do a clean build or switch build config for the sole comfort of a handful of devs at best has irked me for years now. but i have been unsuccessful in convincing the core team this is a bad thing.
ill never vote for adding more stuff in there. but wont stand in the way either as opposing it is seemingly an utter waste of time
22MB of raw source files (yikes! some of these .cpp files are 100kloc, seemingly generated from some XML schema), 2.9MB zipped. So yeah, not exactly small
Yes, I’m using it actively and routinely. So far I could not see any issues with it.
I know nothing about the format itself, other than it is basically “done”?! But it works well, very well, since it is one of the few “universal” formats out there. That being because it is so old (well, actually not but just fresh 20 years; that’s less than half the age of .obj and therelike).
Most vendors of professional software – that is: outside of VFX – out there hesitate to bring in new fancy stuff, that is still changed a lot and in active development (like gLTF and USD) while also probably not even supporting other important stuff, so it’s a plus to have those “well-hung” formats.
Again, I can totally understand that some of you might not be thrilled about having big external dependencies that make the build process a hassle. The mentioned 22mb is a tough call.
What about switching to another library? https://www.assimp.org/ e.g. supports Collada.
I also think while writing it might be a different beast, but at least reading in .dae would be worth much.
The problem with switching to a new library is that it’s hard to ensure feature parity, and it will require for sure some work to get there. One of the key reasons for this announcement is that there is no active support for Collada in Blender for years now.
So at this point, it seems that keeping Collada as a core part of Blender makes little sense. If the community wants to keep supporting this format, this is always an option of course - but preferably as an extension then…
Active support is needed primarily for security reasons (E.g. Fixing security vulnerabilities in the Collada library, and updating the Collada library to use up to date dependencies with security fixes)
There is more than only security. Support is also needed when regular code work in other areas of Blender end up modifying data structure or the APIs used by the Collada I/O code. This adds overhead, and increases the risk of introducing new issues since there is (very) little testing for this format, and the user base is not huge.
The only exportable formats from Blender that support mesh data and materials/textures are .3ds and .dae.
All the other formats are lossy in major ways, or cannot be exported from Blender.
3ds just straight up won’t export meshes that have too many poly (and the limit isn’t high).
I’m not saying this is Blender’s fault, but hopefully Sketchup realize this and offer new import formats soon.
3DS is also triangles only, IIRC. I can’t believe that in this century people even use that format. DXF, also ancient and was never that great to begin with.
Sympathies for the Collada users, but this sounds like a SketchUp issue. They should at least support OBJ.
DXF is the open alternative of a DWG, which originally brought Autodesk to the top of this world, when Autodesk won DWG/DGN war for world leadership with Bentley Microstation.
And it still has no alternatives, for example 3dsmax became an industry standard of a visualization mostly because of a proper DWG/DXF integration.
Low level of a DXF support was one of the most heavy vulnerability of Blender we discovered when tried to switch to it, so we even wrote a custom one.
In general, it is a very demanded technical file format.
Obj file format data type is open, which is good - for example we wrote several parsers and splitters to be able to handle extremely heavy cases (we have to deal with up to 10 gigabytes obj files), but it bakes all the components to the scene origin, which makes it superstable, but also not applicable for proper transferring assemblies, instances, and so one.
So when we substituted DXF with OBJ we made it as twostep function - that imports geometry elements from obj first, and then distributes them according to special file that defines the transforms and copies of all the elements described in OBJ, which is generated by our custom Autocad exporter.
(Blender also doesnot import obj pointclouds, but the solution is on its way)
Sketchup indeed has quite weak file IO, but it didnt changed even after selling it to Trimble.
A lot of programs still use Collada e.g. Google Earth. Also, will it just be hidden like the other import options or will it be removed from the extension library entirely?
Collada is not an extension currently, it’s part of core Blender code, which is a big part of the problem. Having Collada support as a community-maintained extension would be much welcome!
Moving Collada to legacy status has no practical effect on the user, besides the fact that there is no promises about maintenance (bug fixing or adding new features). In practice there has been no maintenance of this format in the past couple of years at the very least, so…
Once it is removed, the whole code will be removed from Blender, so there won’t be any way to use it anymore.
just facing this issue indeed. Using Blender 4.0.2 (Ubuntu 24.04). Thought dae would only be removed with 4.5 or later. Is the missing dae import even related?
Blender 4.0.2 should have the Collada importer and the Blender foundation will not remove it from 4.0.2.
The fact it’s missing from your version of Blender probably means that you installed Blender from your package manager, and they opted to build Blender without Collada support. Downloading Blender from the Blender website may resolve the issue for you.
COLLADA is one of two file formats accepted by The Models Resource. What Makes A Good Submission/The Models Resource - The VG Resource Wiki
Several video game asset ripping tools export only to COLLADA. This is likely because it’s one of the few formats that is fully documented. If support for COLLADA is dropped, people will no longer be able to use Blender to make animations based on video games.
Given the highly questionable legality of using ripped game models in the first place, I’m not sure it would be a bad thing to make it harder to use Blender with ripped game models
But it works well, very well, since it is one of the few “universal” formats out there. That being because it is so old (well, actually not but just fresh 20 years; that’s less than half the age of 