scripting interface design issue
Since the designers of vrml
decided not to provide a "getNode()" method
in the browser scipting interface "because it would allow a script
unrestricted access to almost all of the nodes in the world, severely
limiting a browser's ability to reason about what might and might not be
changing." ( in Annotated VRML2 Reference chp. 126.96.36.199 ), then, why
do they provide it in the EAI?
Probably you would say that it wouldn't have been possible to built an
EAI without a method like this. Fine. But now, we effectively can mess
with unrestricted access to almost all of the nodes in the world, thus
invalidate the original design goal to give the browser flexibility to
scene management and optimisation.
So why not provide it now for internal scripting?
In a related way, I find that method that would access the "parent"
browser ( Netscape in my case ) would be of an invaluable help in giving
more control *from within* vrml scenes to the external world with method
that could be "createHTMLFromString()", "getWindowSize()",
"openWindow()", "getDialog()" etc...
Such methods would open the way to integrated VRML authoring tools
within VRML itself. I think VRML should be self sufficient by itself to
display, create and navigate VRML without any need to rely on HTML at
all. I'm working on something similar to Justin Couch VRML generator but
I'm trying to do it from inside...
Lots of headhakes...
Thanks to anybody that can answer this "why not"