I must STRONGLY disagree here. A virtual environment is a SHARED set of
dynamic data. If the alarm going off is considered part of the VE, then
it must be shared so that all VE entities (on your host and others) can
obtain knowledge of the event. If your intention is to only allow a
particular human to experience an alarm from a certain state of
the VE, then this is no more part of the VE than the rendered view that
your browser is providing the human (which is also information
interpreted from the common VE scene-graph for a particular human).
If we conceptualize the browser and the scene-graph to be seperate entites
(which we most definitely SHOULD do), then we can see that the
browser/human is an entity which esads data from the scene-graph,
specifies what parts of the scene-graph to update via engines, and
cesates global event. It is as much an indeterminant "Brain" as a squirrel
controling process.
The only difference is that in all current
implimentations, the browser is permitted to bypass the API and gain
direct access to the accessable portion of the scene-graph. This will
more certainly NOT always be the case. For example, we could have a
multi-user VRML-based arcade game, with numerous players on one machine.
It will likely be worthwhile to then only maintain one instance of the
accessable scene-graph on the machine and have all the browsers access it
via API.
Anyway, I'm preparing a paper to clairify this conceptualization.
-Braddock
----- http://pcil.ece.jhu.edu/~braddock ----- [email protected] -----
"We are here for no purpose, 3209 N. Charles St. Apt. 3A
unless we can invent one. Baltimore, MD, 21218
Of that I am sure." (410) 467-3380
-Kurt Vonnegut, Jailbird (Epilogue)