X

Bringing 3D to the Web

Backers say a successful 3D programming standard could bring Web developers bold new navigation and rendering possibilities.

Paul Festa Staff Writer, CNET News.com
Paul Festa
covers browser development and Web standards.
Paul Festa
7 min read
Virtual Reality Modeling Language is back from the dead--again.

That's the message coming from this week's Web3D 2002 Symposium (W3D) in Tempe, Ariz., where members of an emerging Web 3D standards group unveiled the first working draft of Extensible 3D (X3D), a successor to VRML. Among its claimed advantages, the new version is designed to work with Extensible Markup Language (XML), which is fast becoming the glue for a wide range of Web documents.

Backers say a successful 3D programming standard could bring Web developers bold new navigation and rendering possibilities that have long flourished in markets such as PC games but have failed to make the leap to the open Web despite years of effort.

"The pump is primed," said Tony Parisi, co-author with Mark Pesce of the original VRML specification and co-editor of the W3D's X3D working group. "Before, we were way too early: The pipes weren't big enough, the computers weren't fast enough, the OS and Web browsers didn't have enough capabilities in them to build upon. In 2002...the stage is set."

Those claiming imminent resurrection for VRML face a tough audience, as the 7-year-old specification has earned its place in Internet history as one of the medium's most overhyped technologies.

Since its ratification by the International Organization for Standardization (ISO) in 1997, VRML (pronounced to rhyme with "thermal") has made little progress toward mainstream use. Though a staple in the computer-aided design (CAD) market and the military--which require the stability of an open standard--sluggish bandwidth, computing power and consumer demand have kept VRML out of more mainstream consumer computing applications, such as games and e-commerce.

Analysts remain divided over whether 3D on the Web will be of much interest to a general audience. Marketers have reported success with 3D retail environments in which products can be rotated and manipulated in three dimensions. Other uses include multiuser games, e-learning applications, data visualization and warehousing, and collaborative design and engineering.

"The market is unknown," said Michael Arrington, an analyst at Penton Digital Media Research in San Francisco. "That was the problem with VRML way back when. Here you have an interesting technology, and there's more fertile ground for these seeds to fall on now that we have faster computers and bandwidth. But what is the content going to be? I don't think anybody has the answer. There's marketing and shopping, sure. But is there a market for walk-through worlds?"

A new horizon for 3D
The hopes of X3D backers are riding on developments that have created a more conducive environment for 3D on the Web, including improved bandwidth and computing power, significant improvements in the underlying schema, and support from other standards bodies.

Parisi, who recently launched Media Machines in San Francisco to provide tools and consulting to Web sites using X3D content, said the new 3D standard improves on VRML's unwieldy size and structure.

VRML's roundabout path

1994-1995 San Francisco software engineers Tony Parisi and Mark Pesce spearhead an effort to standardize 3D on the Web.

1996 The VRML Consortium is formed.

1997 The ISO ratifies VRML 97.

1998 3D on the Web founders. Silicon Graphics shutters its Cosmo unit. Microsoft shelves its Chromeffects initiative.

1999 The VRML Consortium, in an effort to create a "commercially relevant" successor to VRML, regroups as the Web3D Consortium.

2000 The X3D working group is formed.

2002 An X3D draft is released.

X3D comes in a modular structure built around a small playback engine, in contrast to its predecessor. While VRML weighed in at a cumbersome 2.5MB, the smallest part of the segmented X3D specification is a comparatively lithe 300KB.

The specification also describes three profiles, or common configurations, of X3D modules. The first, Interchange, is responsible for geometry and animation and for exchanging data between authoring tools. The next level, Interactive, adds to the Interchange profile interaction with elements in a 3D page, such as mouse-overs and click-throughs.

The third profile, Extensibility, lets 3D authors create X3D components and link their applications to databases and other outside sources of information.

Those familiar with the proposed standard said X3D provides a more complete base than VRML that developers can extend without fragmenting the core specification--a problem that stymied VRML, said analyst Jon Peddie, a principal with Jon Peddie Research.

Other forces that could help X3D include growing support from other standards groups. Notably, the influential Moving Picture Experts Group (MPEG) agreed to include X3D in MPEG-4, an emerging standard for digital video and audio. Since the use of MPEG-4 has been held up by a licensing quagmire, however, that road to the specification's success remains uncertain.

Backers point to a third catalyst that might finally ignite the 3D market: the use of advanced graphics in operating systems.

"We're still constructing the applications that use 3D all day every day," said W3D Consortium President Neil Trevett, a vice president at graphics-card maker 3DLabs. "You'll start to see that kind of application when the OS begins to use elements of advanced graphics, including 3D. It's just like moving to a graphical user interface from the old command-line prompts; after the operating systems did that, the applications followed."

Trevett said that process had already begun, albeit at the early stages. For example, Microsoft's PowerPoint application running on the company's Windows XP operating system is "the first time a mainstream corporate application begins to use advanced graphics," Trevett said. "Pretty soon, even your desktop is going to be graphics-accelerated."

The Microsoft card
X3D backers point to Microsoft, with its command of the Web browser market, as a potential spoiler. Should the software giant provide built-in support for X3D in Internet Explorer, it would pave the way for broad acceptance by eliminating the need for an extra download. Conversely, its refusal would only hurt the budding standard.

Microsoft declined to comment. Close observers of the 3D market said the company would hold its cards close to its vest with respect to X3D while it waited for strategic factors to clarify themselves.

"The reason you can't get a straight answer out of Microsoft is that it still struggles with the question of supporting open standards vs. using proprietary software," said one analyst who asked not to be identified. "They are an obstacle to this happening because they're trying to get a proprietary edge on it. They're having the same problem with MPEG-4. They won't support an open standard until they're forced by the market to do it."

Microsoft has a troubled past when it comes to Web graphics technologies. Its Chromeffects initiative to bring DirectX-style graphics to Web pages was shelved in 1998. Bits and pieces of it were later released in the company's Internet Explorer Web browser.

Like any open standard, X3D faces competition from proprietary alternatives. Many competitors have handily beat X3D to market, unhindered by the perennial, painstaking search for consensus that lies at the heart of the standards process.

In the face of these swifter competitors, X3D backers say, the standardized approach has some built-in advantages.

"Standards offer the promise that you have multiple vendors creating tools and reusable content," Parisi said. "If done right, there's a whole economy that can spring up around it. Look what exploded around HTML. That happened because everyone could use it and nobody owned it. It's the same with XML."

That said, the W3D Consortium acknowledges that some 3D applications--multi-player games, for example--are likely to remain the province of proprietary software.

Silicon Graphics, which contributed the original VRML code from its Open Inventor 3D visualization product, has largely abandoned the Web 3D market in favor of a system of closed "visual area networks" for sharing 3D models. The company shuttered its 3D subsidiary, Cosmo, in 1998.

But SGI, which still supports VRML in some products, hasn't ruled out giving the Internet another chance.

"VRML and the structure of the Web were not set up to interact with large models," said Janet Matsuda, director of marketing at SGI. "But we see (visual area networks) going toward the Internet."

The Web3D Consortium is itself a member of the World Wide Web Consortium (W3C), the premier Web standards body, and is touting X3D's close integration with W3C technologies.

W3C recommendations employed by X3D include the Document Object Model, which lets authors manipulate elements of an XML document using scripts; XML itself; and the W3C's recommendation for handling two-dimensional graphics on the Web, Scalable Vector Graphics.

Looking ahead, the group plans to employ the Synchronized Multimedia Integration Language, or SMIL, for synchronizing 3D elements with other media types.

Now that the Web3D Consortium has issued its draft, it will solicit feedback before submitting the specification to the ISO in August. An ISO standardization could come as early as next year.

One analyst predicted success for X3D, saying the technology and demand for it had finally come up to speed.

"There is a genuine pent-up demand for this stuff," analyst Peddie said. "What's out there now is primitive and not very satisfying, but that's partly because the people putting them out have to develop them from scratch. If they could get the same level of support and functionality with 3D that they do with XML, you would see amazing new developments."