DTD issues- NEWS FLASH !!!
A vote on Monday 3/13 by the reached consensus (77% in favor) to adopt the 'Compromise DTD' as worked out since Monterey. While an X3D DTD (the XML 'tag-key' file which determines validity of X3D documents) is not a final recommendation, it does somewhat shape the nature of the beast.
In this case, tag element proliferation makes X3D slightly more bloated and difficult to author by hand, but guarentees a parameterized, valid, self-described scene-graph. This means both portability (no apriori knowledge needed to parse the scene) and extensibility (can explicitly define new fields and field types) for X3D.
When we build 3D scenes for the web in VRML, all nodes have a a set of various fields to them which describe the parameters of the node and provide 'hooks' for events to be routed in and out of that node. In VRML 97, all of these are defined in the spec.
With X3D, the game is a bit different as new node types can be dynamically added to the scene and must be handled properly. In theory and with XML, this extensibility requires nodes and field to be semantically transparent (explicit tag elements for both in the DTD). Whether this current Compromise DTD can guarentee a DOM compliant Script Authoring Interface (SAI) and keep file size small (Working Group tests show a size increase from 0.5%-41.0%) remains to be resolved before SIGGRAPH 2000..
Back In Monterey...
The debate on how to define the DTD worked late into the night as each side edged toward understanding...
Wrapper Tags or no Wrapper Tags?
Shout Interactive got creative and broke the week's tension and hectic pace with a little pugilistic pleasure... in 'The Fray in Monterey' boxing game bwtn. Chris and Don was a great hit and I'm still moved to laughter by these images!
While Chris Marrin and others maintain that wrapper tags are necessary for unambiguously representing multiple SFNode & MFNode fields, Don Brutzman and others proposed their tested work on the X3D_draft.dtd which avoids potential naming collisions and keeps the code compact. While this does require the parser to perform some extra processing (apriori knowledge of scene-graph semantics), Don et al. argued that that for VRML 97 compatibilty, this processing would happen anyway...
The Proof is in the Pudding as they say!
Questions, Comments, and Suggestions about this site can be emailed to: [webmaster]
VirtuWorlds, VirtuWorld, VirtuPortal, and 3DEZine are Trademarks of VIRTUWORLDS LLC.
No unauthorized uses are permitted.
© VirtuWorlds 2000