A good point. I would say that we really want to tease out the "basic"
kinds of objects, the ones that really define a space. Sort of like the
way that X defines a "window" right in the basics, but doesn't really
restrict what that means. I'm not sure what these basics might be; a
"scene" is probably one (although I'm not certain), a "thing" another
(most normal objects would subclass this), and maybe a "port" another
(if we decide that we need to define the characteristics of navigation
at the language level). Quite possibly, we need only one truly basic
meta-class, akin to the "Core" class in X -- pretty much all X objects
are subclasses of "Core". This issue bears a little thought...
(There's nothing wrong with having several levels of object
definition, BTW. We might define just the *concept* of the
object-orientedness in the language, add to that a standard small
collection of core classes, and encourage other groups to come up with
various standard object collections falling within that standard. The
more work we can foist on others later, the better...)
>So, every VRML file that "inlines"
>another object must guarantee it'll have that object (or references
>an object at another site guaranteed to have it) so that people without
>caches and CD-ROM's can see everything, but those with local caches
>and CDROM's also win.
is a nice concise statement of the point I've been trying to make --
it's good to widely distribute the objects, but the ultimate
responsibility should fall on the site providing each scene...
Random Quote du Jour:
"[By the way, I've always wondered about this: Before he became Caspar
the Friendly Ghost, was he Caspar the Friendly Terminally Ill Child,
Caspar the Friendly About To Be Brutally Murdered Child, or Caspar the
Friendly Horribly Mangled in a Terrible Accident Child?]"
-- Dr. Foo