The standards group responsible for the graphics exchange standard realizes content exchange is a huge challenge and that so far Collada implementations are falling short on. Commentary by Kathleen Maher.
By Kathleen Maher
The Khronos Group has published an open letter to the content creation community about its Collada exchange format. You can read all about it (at GraphicSpeak or the Khronos web site) if you haven’t already. Essentially the letter is an admission that the problem of content exchange is a huge challenge and in the process of trying to solve many problems in the game development pipeline, Collada implementations are falling short on some of the basics.
As the letter says, “We realize that for many in the industry Collada is not currently providing reliable import/export capability from key authoring packages such as 3ds Max, Maya and Blender—significantly diminishing its value in enabling authoring pipelines.” At the same time, Khronos is resolute in its belief that the industry needs an open, non-proprietary file format like Collada—for people to exchange, deploy and archive 3D assets in a format not controlled by any single company. Khronos says Collada needs the increased participation of the 3D community to make it work. Khronos has proposed the establishment of a Collada Forum, freely open to all, focused on enhancing and building out the open source OpenCollada import/ export pipeline to create a community-maintained Collada that will be rock solid.
Khronos is reaching out through the open letter to engage the energy of the 3D community to help identify and clean up the bugs and implementation issues that are holding back Collada adoption. Khronos hosted a meet-up at Game Developers Conference this year to get the project started. It was well attended and there were quite a few people who professed commitment to the goals of Collada. Khronos has pledged to do their part with funding to fix currently recognized OpenCollada bugs and to simplify the publicly available conformance test suite and publish the results for 3ds Max, Maya, and Blender import/export to create a baseline for directed improvement.
Collada is a large and complex specification, and Khronos realizes that many pipelines in the industry today would be more than adequately served with a reliable subset for model and character animation—consequently Khronos has proposed to focus OpenCollada bug fixing efforts on an agreed subset of Collada 1.4 as a first step. They’re looking for input and feedback on this approach and will be opening up a forum for people to get involved.
Collada is at an awkward stage. There are important projects that have been built using the format including Google Earth and SketchUp. Apple supports Collada in Mac OS Lion and for its iBooks. And there are pockets of Collada support in the CAD industry. It is in process of becoming an ISO standard to be used by the architectural industry among others. In addition, Collada has a whole new wave of opportunity opening up—to enable rapid authoring of WebGL content. There are many people out there who depend on Collada, they have committed to it and they would like to see Collada continue to be developed so that it really can become a useful exchange tool.
Khronos President Neil Trevett told us that they’ve come to realize that while the larger vision for Collada is ambitious the industry needs to get the basics implemented reliably first —and the step Khronos is taking is really an attempt to get back to those basics. The focus is on the most immediate needs, exchanging 3D models for everyday pipelines in widespread use, and to then build out to more industries and workflows from there.
The Khronos Organization is not washing their hands of responsibility for Collada development; they are committed to maintaining the Collada working group within Khronos. It will be a forum open to companies wishing to develop and evolve the Collada specification in a proven cooperative environment.
Khronos is taking the further step to encourage more active support from the 3D community and from the people who will be relying on the format in their everyday work.
Time to regroup
Back in the misty past when Collada was first proposed, Khronos saw sellout attendance for meetings. The ability to exchange data between programs remains a real need but a significant challenge. Autodesk, as we all know, is the 300 lb. gorilla and it has its own exchange format in FBX. Autodesk has pledged to be a good citizen in the 3D community and on the whole it has been, but some of the Autodesk Collada import export bugs in its implementations have fragmented the Collada community, and there are many companies that would be considerably more comfortable if there were a workable open alternative to FBX that can’t be unilaterally changed or withdrawn by any single company. In addition, as the evolution of Collada has demonstrated, there are applications that are enabled with an open exchange format for 3D content, such as the fast evolving environment of the cloud, which is a great opportunity for something like Collada, which could let people exchange content without really worrying about formats and compatibility.
Part of the problem for Collada is not that it hasn’t been accepted; it’s more a matter of having been accepted by so many different companies that there is fragmentation. Alembic, another attempt to share content, has grown up in the film industry. The existence of Alembic and Collada demonstrates the need for some kind of format.
Intel has tried to also find some common ground for 3D exchange. It really is time to regroup. Some of the people that most need a 3D exchange format don’t even know about the work others are doing. Khronos is right to reach out and try and create a community that can come together to solve the big problems of communication among applications and content creators. It might not work, but it’s worth trying.