If I might give you an advice: Restrict yourself. It’s almost impossible to support everything collada allows. So pick the stuff you need and ignore the rest.
Yea, that seems to be the easiest way, to arbitrarily pick the concepts you need and use them without caring about the rest. This has one major disadvantage - it makes your code error prone, and what is even more frustrating, you can’t solve a problem when it appears, because you are not familiar enough with the underlying architecture (in this case COLLADA). So I want to at least be familiar with the Basics, the generic structure of the .dae file.
It’s hard (for me) to learn the basics of COLLADA, because the specification is rather table-oriented than descriptive. For example, I still don’t know precisely how the scene graph is being organized, after reading the entire chapter about <node>. Or how all that referencing and source-ing works. That’s something you can see best in an example or tutorial, and I end up reading the example .dae-s in the DOM-viewer to figure out how its done It’s definitely not the best way to learn collada and I’m sure I missed somewhere that chunk of documentation which describes it in a consecutive, descriptive and most important, understandable for us poor lamers way
I suspect the book might be a choice, but its rather expensive and I don’t want to buy it before I’m sure I want to use COLLADA.