I completely agree. Specification of sound sources can easily be done using
the extensibility features of VRML, allowing browser writers to experiment
with different specifications (in their copious spare time-- HAH!).
Lets rephrase this discussion as:
"I'm a browser writer, and my browser is going to support this on an
experimental basis:
Sound {
fields [ .... ]
...
}
Is anybody else experimenting with sound, and, if so, how are you thinking of
doing it?"
In my opinion, the right process for extending the VRML spec should be:
1. Discussion of desired features. Sample implementations tried BY AUTHORS
(not browser-creator geeks, please). Re-specification and re-implementation
until the feature is solid.
2a. Proposal to add as a standard, "must-be-supported" part of the spec.
2b. Authors LOVE the feature and demand all browsers support it.
(these go on at the same time)