Sending Info
The Bagel-Meistro (cbentzel@media.mit.edu)
Tue, 14 Jun 94 12:06:34 -0400
Hi. It's time for me to add to the conversation.
Currently there seems to be a debate of using a library of
previously created shapes, or to send the info describing
the construction of the object over the network. It is not
inconceivable to do both.
Using a portable, PHIGS or PHIGS+ sort of language, one
could send out either a) a description of how to compose the
shape from very primitive polygons, or b) send a name of
a previously created and publicly created object. However,
by sending the name, this just reduces the amount of info
transmitted: it still could be defined using the same
primitive polygons or built out of increasingly complex
objects. The host computer would still have to process the
same amount of information, but the drag from the amount
of time the information takes to go from server to client
would be reduced dramatically at sites that just sent the
names of library objects, instead of the libraries
themselves.
The other thing that could be used with this method is to
use a statement similar to <IMG SRC= url ALT= foo> in html.
In the vrml language, the SRC would be equivalent to the
name of the object, which could be culled from the clients
personal libraries, and if the client does not possess that
object, would be sent the description like one is sent text
if their platform does not support images using html. Just
some thoughts.
Bentzel