Where Does A 3D Standard Start?

Noting a news page at Wired quoting Sutor at IBM about the need for virtual world standards... sigh. Where to start.... The problems are not disimilar to OMG. Moving an object among different platforms worked pretty badly until the technical writers taught the programmers markup. Then the great AHA moment of 'it's the data, dummy' happened with XML.

But that's just a format.

Real-time objects aren't just data. They have the polarities of rendering and behavioral fidelity in a system where local real-time (what you see on the screen) is in its simplest form, frame rate. Drop below about 15fps and you might as well be watching a Powerpoint slide show. To solve that, X3D specifies an object-model in the standard. That makes interoperability of real-time 3D content nearly doable with some wrinkles being worked out where the standard is silent or ambiguous. This is not an easy kind of standard to write and those proposing ever more complex features based on experience to a single platform's capabilities are tredding water very naively.

The impact of all the marvelous AI and other bits Bob talks about have to be measured in terms of cost to FPS. The standards start there, not in Linden Labs or IBM. Otherwise, the people just entering this market amazed at their LL-hosted worlds will propose just another proprietary framework with a physics engine hosted on a server farm and replicated for a fee.

That's not an open standard. We already have those: see X3D, Collada, etc. For some reason, those who yell loud and long about Microsoft and OpenDoc don't seem to understand the same rules for open unencumbered standards have to apply to 3D as well. Why?

I have my suspicions, but the IBM standards reps will have to answer those questions. Bob?