Why is there so little usage of XML and X3D on the ‘visible Web’?
Why is it that 3D web pages didn’t become ubiquitous and commonplace, as the hype predicted 5-10 years ago. Back then, people were touting that it would only be a matter of time before everyone started building 3D web sites just like they were building HTML pages. Is it that 3D on the web failed? Or is it that many of us didn’‘t really understand that the Web is a much bigger and more diverse place than HTML pages? X3D, particularly in it’s XML incarnation, is actually growing very very rapidly on the web. But it’s not growing as HTML pages - it is growing as real XML-based applications that demand serious technical chops to develop. The applications are project specific, and don’t simply scale up for “consumers” to develop on their own.
To give a little data for why 3D or any application on the web that requires serious technical chops doesn’t scale up like HTML (again, intelligence doesn’t scale), see http://technology.guardian.co.uk/weekly/story/0,,1823959,00.html,
Note that with the requirements for messy protocol code and the ability to think in a multi-dimensional application (XYZ + Time(User- intent + Sentient Environment), you have to have professional 3D graphics skills. This isn’t the world of 2D vector or pngs although you have to have these too (textures and 2D layers). Given the figures in that article cited above, one begins to understand why comparing the HTML authoring and the X3D authoring communities concludes that all they have in common is XML skills and that is trivial in the sense it is assumed. Basic HTML could be picked up rapidly and just a few element types do the job. Hypertext is easy. As you add media types, it gets harder. X3D is bigger, requires geometric and animation expertise to do the simpler tasks, and serious programming knowledge to do industrial work. All of that is needed before you even get to semantic domain technology (what are the rules of emergency responders in a situation of type zed). This is why tool support and libraries are so critical. Fortunately, the consortium has stepped and gathered support across different organizations and companies to get X3D into export packages and authoring suites. Companies like Media Machines have made their brower code open source so the ‘softies have a basis to get started. Companies like Yumetech have stayed the course with open source libraries for Java. The US Navy at the Naval Postgraduate School has done herculean work to provide libraries, guidance, projects, research and leadership. There are many others. Comparing them to web page companies is like comparing Juliard to your local technical college.
This is not easy work. The artistic worlds are hard enough but the serious apps require serious effort. In the breezy land of ‘make a company fast; flip it and move on’, this is too hard. For the companies and individuals with depth and commitment, it is a satisfying challenge and there is increasingly good money to be made because there are RFPs on the street that cite X3D and budgets to back them up. You have to compete against the likes of David Colleen at Planet9 and that won’t be easy. It will help if you come from a university that made VRML coursework part of the curriculum and those tend to be in Europe where VRML never went off the radar. It is tougher in the US where commercial authoring packages, not languages, tend to dominate 3D coursework.
You don’t use X3D because you want to build the next WoW game site. You can but you won’t. You use X3D because you have a real time web capable application that needs royalty free technology using scene graph technology that is reliable and repurposable for a long time regardless of the fortunes of the company that provides your tools or browsers. It isn’t perfect because there is still work to do on the standards, but because that has been going on below the radar with a smaller dedicated team of companies and individuals, it didn’t suffer the hysteresis that has ensnared too many W3C efforts. There are advantages to being considered ‘dead’. It keeps the work focused on staying alive.
Nothing worth having is easy to get. The more complex the application, the fewer the types of users. The more types of users there are, the less complex are the objects. So any complex 3D app has a smaller set of users and any one that a lot of people use is necessarily shallow (in terms of the different things one can do with it). HTML is in the sweet spot of hypertext applications. X3D is in the sweet spot of real time 3D applications, but these are not comparable beyond that.
Real time 3D challenges web mythology at every level.