What you are talking about are the ‘genre’ of 3D as Eric Maranne puts it. It is important to distinguish these in the market and in your products. What you call ‘sims’, is also referred to as ‘virtual reality’ and that is a better term because it opens up more content potential. You may not be simulating anything. One common feature I find among people who are good at that genre is they all seem to have a background in theatre. It is theatre tech in pixels. It may not be a metaphor or it might.
Visualization is distinct in that it really is strictly metaphorical and that is one market that tends to run out of steam because a metaphor is only as useful as it is easily comprehended and far too many of the common practice visualizations take experience to comprehend even if they are great eye candy. But the common functions for transformation, rotation and zooming work. The distinguishing quality of 3D is whether it is static or real-time. Real-time does not mean ‘animated’. It helps to understand the difference between say a ‘frame-based’ animation (think datasets for interpolators) and script-generated motion (think motion and direction vectors being determined by a script that uses rays to determine the classes of objects in proximity). In X3D, mastery of the sensors is the first level to developing convincing scenarios because they principally determine what goes on vis a vis the user’s position, not among objects in the scene.
I’m not sure what you mean about ‘world views’ and X3D being a problem. That is actually one of its strengths if I understand you. X3D works great. It is not just VRML. BTW, a good book is coming out in April from Don Brutzman and Leonard Daly.
As with HTML, X3D is a hypermedia wrapper language. Unlike HTML, it is a bit lower level. View source helps but not nearly as much as it does with HTML. Cutting and pasting out of transforms can lead to bizarre results because of accumulations in nested 3D coordinate systems such as scaling. To create presentations, you have to think about time, location and orientation in ways that are unfamiliar to page builders. You may also, as in MU ( the multi-user variation for VR), have to consider what the presentation is with multiple users roaming it at the same time interacting with the environment. What happens if proximity sensors are live and trigger behaviors when you are attached to a camera in motion? Do you animate the camera at the same time? Where do you go if a viewpoint is unbound (it pops the stack, but what was the last viewpoint selected and does that shock the monkey)? How do you route an event among different worlds that are composited into a main world (the most common way to build)? Is it better to use inlines (easy) or protos (flexible and the real component level for builders)? Protos are better because inlines are event-opaque.
What we learned over the last decade of working with 3D on the Web is that the simpler languages (eg, 3DML) are easy to start but not powerful enough once past basic scenes. X3D/VRML is harder but at the sweet spot of power and ease. There is no free lunch here. More can be done with high level languages and syntax transforms. Ajax3D works but it always did.
It is a real non-linear system. It is easy to think ‘this is just games’ but that is just a genre and as with any genre, it has its own constraints (why are so many games in closed spaces even if they don’t appear to be closed, for example)? Each genre has its own vocabulary. The amount of academic BS there is very deep and mostly useless. Because the basics are basic, all of the meaning is in the expression.
BTW: want a good cheap editor? Flux Studio (Media Machines) is free for personal use. Blender is free but has a steep learning curve. OTOH, it can do anything. If you want to play with Boolean carving, check out the ISB demo from Parallel Graphics. The Flux viewer is open source. The Xj3D libraries (Java) are open source.
makes sophisticated 2D possible. Apple proves this daily - and most of the core
changes occuring on the Linux and even MS side visually are due to the
integration of 3D processes into the daily workflow. Sure, many of them are
just "effects", but its remarkable how effective those effects can be
for making metaphors believable.