One 3D Format to Rule Them All

prostoalex writes "Three-dimensional graphics for the Web always seemed like a great concept that's not there yet. Five years ago many publications saw a great future in 3D-Web, but somehow things just haven't been moving in that direction. Apparently, the status quo is not making companies in this field happy and so the big guys, including Intel, Macromedia, AutoDesk, EDS et al. formed a 3D CAD working group. They claim that 'the need for a common 3D format becomes clear in a simple perusal of the Web, where the volume of 3D content is minuscule -- well under 1 percent.' The article is published in the latest issue of Intel Developer Update magazine, which is also available as a PDF."

Comments

I'll tell you why 3D web content is so scarce

[Featureless]

And I should know, because I do a lot of work producing it. It is by no means a lonely field, but there are relatively few people who do it.

Why put anything on the web? It's relatively cheap compared to printing flyers or magazines or novels, and there is really no parallel for instantly delivering interactive media (I've done dozens of web-based games, 3D and 2D using all of the technologies you've heard of and I'm sure several you haven't). But really, why?

You see, 3D is complicated. I've regularly had to participate in hiring of modelers and artists capable of collaborating producing good, efficient 3D art on a deadline, and real skill in this field is still rare. I know the tools, and I've watched them work, and I see why. It takes a unique blend of manual dexterity, artistic ability, spatial skills, math, and geekdom - especially the last, because you have to be a geek to keep up with the tools and the issues, which are heinous. That law about the more special-purpose and expensive a piece of software is, the worse it is, applies to 3D tools in spades. There are so many bizarre little problems.

Last but not least, most of the widely used 3D authoring systems are, or have historically been, very very expensive - $5,000 - $10,000 - $25,000 is not an unusual amount to spend just on software. There are cheaper tools, but remember, you have to interoperate with web middleware, and pretty much everything just imports from 3D Studio Max. And then, what's your presentation platform? VRML (ech)? Shockwave 8.5 (~$1,000)? There are others... my point is that most of these cost money too. Pre-rendering to Flash is the cheapest and actually very attractive, but then you don't get anything in real time and it's really just a clever trick for making a canned animation.

The net result is that there are very few hobbyists producing 3D for the web - games or anything else. And then we have companies. So why would companies want to produce 3D content when they get almost as much oomph with good 2D technology (or just plain graphics) without the significant costs, and endless technology headaches? That's because even with the best middleware, you might see 10-25% of your users have some kind of 3D hardware/software related problem... old video drivers bunging up D3D which bunges up whatever your middleware is, weird budget 3D cards, software mode, etc... Most businesses just want to spend the least amount of money to reach the largest possible audience. And that was true before the drive to produce any kind of non-ecommerce-related commercial web content at all pretty much dried up.

There are still a few people left who we haven't eliminated who, for whatever reason, it makes sense to produce 3D for the web. Product demos, promotional games, and the rest. Believe me, competing for their business is far from easy. ;)

Don't get me wrong; I welcome better tools and better standards. There might be a niche for simple object inspection or static environment presentation ala a not-totally-braindead-VRML. But it's really not a big deal at all. Most of the real issues to do with web 3D are on the OS side of the equation - uniformity of hardware, APIs, cross-platform issues, etc. IOW just "Stability" and "Reliability." It's really, really hard to deliver 3D content to a wide PC audience even without the massive additional headaches of the web. At the end of the day, I think web 3D will come into its own when we collectively find it easy to author 3D the way we author text-and-graphics websites now - in other words, maybe never. Until then, it will remain a specialized niche which is (these days) reasonably well served by the existing toolmakers and not really susceptible to wide-ranging standards due to the divergent needs of the participants.

Uh oh, Web3D Consortium

[Animats]

There are really two completely different projects here. One is X3D, which is basically VRML 97 in XML syntax. The other is Hoops 3D, which is an binary interchange format for CAD documents. These are completely different. The two organizations are having some meetings with each other.

X3D is definitely a solution looking for a problem. The Web3D crowd basically killed VRML by announcing their "new, improved, XML-based solution", used by nobody, supported by nobody, and with very little active work. If you get their SDK disks, it's mostly old VRML stuff and old Java3D stuff. This was sad, because it happened just about when hardware got good enough to do VRML properly. VRML over broadband with a current-generation OpenGL board works quite well.

I haven't looked much at HOOPS, but it's a reasonable idea. IGES, the old FORTRAN-based interchange format (80 column lines, no less) is a bit dated.