Thanks for the feedback, after considering opencollada and all its great functionality, we've decided to stick with collada-dom for now. Although it has problems (what library doesn't?), it gives the user as much control of the parsing process as possible. We're also focused on parsing robotics relevant data only, so we're not much interested in extracting and using the correct effects/materials/animations.
Now that I have developer access to collada-dom, can I just start committing my patches? Do I have to send them to someone first for review? What about releasing minor collada-dom versions with it?
I've also read elsewhere that collada-dom was mostly auto-generated from the spec. This is one of the reasons why it is attractive for us. Just curious, where can we find the program for this generation?