Am 24.01.2007 um 19:05 schrieb Peter Amstutz: > I agree. I have the book they published describing the COLLADA spec, > and intend to base the VOS 3D data models on COLLADA wherever it makes > sense, including physics parameters. > [...]
I guess I'll have to chime in at this point... I only took a quick glance at COLLADA, but it struck me again as a format primarily intended to model the *appearance* of a scene, rathern than its semantics. In that it isn't much different from, say, X3D, only that it may have more up-to-date features such as physics, scripting, and shaders. So once again: A shaded ball on a shaded cylinder being transformed to bend slightly along the Y-axis is just that, it is NOT a tree bending in the wind, even if it *looks* like one. If you ever expect any kind of autonomous machine interaction with a VOS world, please design the 3D data model so that it can co-exist with a semantic data model. Even better, make the semantic entity a first-class memeber of the world, and put the appearance in a child (or even external reference), i.e. world ---member--> tree ---appearance--> (cylinder, sphere) world ---member--> house ---appearance--> (box, extruded triangle) much better than world ---member---> (cylinder, sphere) world ---member--> (box, extruded triangle) and expecting anybody else than a human with a visualization program to ever get the meaning you want to convey. Of course, if you saw "Matrix" a few time too often, there is no better way to get rid of an "agent" than placing it in a world it cannot possibly make sense of :-) Touching another topic, this kind of semantic design also gives you a selection criterion for scene querying and caching, reducing bandwidth and memory overhead. If your client is only interested in trees but not houses, why should it download the complete appearance definition of the house? Also, if it has no 3D display capability, it might want to download only the member metadata, and possible a different form of "appearance" such as a text or 2D icon. Regards, Karsten Otto (kao) _______________________________________________ vos-d mailing list vos-d@interreality.org http://www.interreality.org/cgi-bin/mailman/listinfo/vos-d
