In this section our purpose is to demonstrate the use of an interpreted language to control aspects of a 3D interactive space through creating, instancing, placing, and modifying VRML nodes.
Our first example is simply the IVL interpreter on a web page[6]. This page has an HTML textarea in which the client can type in IVL commands; the resulting scene graph is returned to the client's VRML browser.
The second application is a little more ambitious. We would like to view VRML scenes as part of an integrated, seamless space which one could travel through while maintaining relation to the previous space. Currently in VRML, clicking on a hyperlink replaces the current scene with a new one. This leads to an effect sort of like the transporter in Star Trek; you are instantaneously ``beamed'' from one scene to the other without experiencing the distance in between. With this paradigm, the spaces are not associated, rather they are completely independent with no sense of attachment to each other. In the virtual world we wish to create, there is a sense of unity since the spaces all have enforced physical relationships to each other. So, I could walk out of my front door and be on someone else's mountaintop for example, but I would be able to turn around and see my house. Note that this does not preclude the use of hyperlinks to jump to other parts of the world---we have just added the relationship between spaces which was previously lacking.
In order to get the most out of these example applications, some browser modifications are necessary (although they do work with stock browsers). While our original application was based on the Inventor SceneViewer application, we have begun to migrate the code into the public-domain SDSC browser so that people can download our modifications and try them.