X

Web 3D: Back to the drawing board

The year 1999 tried the patience of advocates for 3D Web graphics, with definitive breakthroughs in both technology and demand sorely lacking.

Paul Festa Staff Writer, CNET News.com
Paul Festa
covers browser development and Web standards.
Paul Festa
7 min read
Among the Internet's many visionaries, few have shown as much patience and persistence in the face of adversity as those behind 3D Web graphics technology.

This year was another that tried those advocates' patience, with definitive breakthroughs in technology and greater demand for 3D Web graphics sorely lacking. Instead, 1999 brought plenty of frustration to graphics heavyweight Microsoft, elevated the efforts of Sun and a few start-ups, and introduced streaming giant RealNetworks to the arena.

Meanwhile, an effort to 1999: The year in technologypromulgate a new industry standard for 3D on the Web plugged along despite the withdrawal of a promised open source software contribution.

Advocates describe a fundamental problem with 3D Web content and technology vs. consumer demand: Although 3D graphics constitute big business for the PC in both the industrial graphics and gaming markets, and although PCs have become increasingly capable of rendering high-end graphics, no one has demonstrated a similarly compelling business model for 3D graphics on the bandwidth-constrained Web.

The Web's limitations coupled with 3D's enormous computing firepower demands caused Web graphics pioneer Silicon Graphics (SGI) to abandon the business years ago.

"The Web is great for viewing some models, but it's not fast enough and can't handle the types of data and kinds of rates that we're talking about," said Chris Insinger, product line manager for SGI. The company sells hardware and software for industrial design and computer animation graphics, such as those used in the building of aircraft or in the making of movies such as "Toy Story."

On the Web, 3D made incremental progress in 1999, with new technologies that have the potential to make both high-definition gaming graphics and business applications--such as design collaboration and data mining visualization--viable businesses online, according to backers.

A key effort to bring PC multimedia effects to the Web stumbled badly at the end of last year with Microsoft's decision to shelve its Chromeffects initiative. Chromeffects would have given Web developers the ability to access Microsoft's popular DirectX technology, a set of low-level graphics application programming interfaces (APIs) used in popular PC games.

The goal of Chromeffects was to make it possible for Web developers to access DirectX, particularly its Direct3D graphics component, using comparatively simple Web programming languages rather than more complex computer programming languages. But Chromeffects encountered fierce criticism from developers who saw its preview, and Microsoft quietly abandoned the project amid disarray in its 3D efforts.

Chromeffects' pieces were shipped with the Internet Explorer 5 browser, but without Microsoft's active promotion, advocacy and training, the product was pronounced dead on arrival by developers.

Microsoft did not return calls seeking comment but previously has said its Chromeffects efforts paid off in the limited release with IE 5.

Microsoft's strategy of making Direct3D accessible via higher-level programming languages lives on in a start-up launched by a former Microsoft employee and inventor of DirectX. For Alex St. John, founder and chief executive of Redmond, Wash.-based start-up WildTangent, 1999 was a year of acquisition, funding and the advancement of two products for creating 3D graphics on the Web.

WildTangent launched a trial version of its tool for 3D content creation in June, and in September it announced plans for a more advanced tool for game development. That tool was developed using technology acquired with Eclipse Entertainment.

St. John says WildTangent will surmount the business model and consumer demand problems by partnering with "major portals" on gaming and mapping applications, though he declined to elaborate on negotiations. WildTangent has partnered with Compaq Computer to add multimedia graphics capabilities to Compaq's Internet kiosks.

"Nobody wants 3D on the Web, because it's useless outside of any context," St. John said. "It has to have an application meaningful to the real world, and we're really focused on using it where it adds value. There's an outrageous demand for 3D, for instance, in the consumer space for games, and we deliver those over the Internet."

WildTangent is in the midst of securing its second round of funding from "prominent Silicon Valley investors," St. John said. First-round investors included New Millennium Venture Partners, retired Microsoft employee Cameron Myrhvold (brother to Microsoft veteran Nathan Myrhvold, currently on leave) and Steve Victorino, who sold 411.com to Yahoo.

As Microsoft's efforts stumble and WildTangent's rev up, Sun Microsystems claims to be solidifying its Web 3D efforts with what it termed "tremendous momentum" for its Java3D technology.

Like Chromeffects and WildTangent's drivers, Java3D is a more easily accessible layer between Web programmers and the nitty-gritty of 3D programming, which involves the painstaking creation of polygon diagrams to be filled with textures, colors and movement.

Sun so far has positioned Java3D in narrow markets where 3D models are already a staple of day-to-day industrial tasks, but it envisions other applications, such as customer service or sales uses in e-commerce sites.

Another company eyeing 3D's potential for e-commerce is streaming media giant RealNetworks, which earlier this month demonstrated 3D capabilities added to its streaming products through a partnership with WebGlide.

In the year since Java3D first shipped, Sun says the API has seen 50,000 downloads. Implementers so far include the National Center for Supercomputing, which uses Java3D for a visualization of galaxies that lets two people look at and manipulate the same 3D model over the Web.

"The collaborative aspect of 3D on the Web is very important," said Kirk Mosher, group manager for technical Java at Sun. "To have two people working together on 3D is one of the new, new things."

Java3D is an implementation of a new specification making its way through the Web 3D Consortium, a group of companies collaborating on an industry standard for 3D. This was a pivotal year for the W3D Consortium; it was born from the ashes of the Virtual Reality Modeling Language (VRML) Consortium and transformed its product from VRML to Extensible 3D, or X3D.

Born at the dawn of the Web as the brainchild of engineers Mark Pesce and Tony Parisi, VRML reached the height of its fame--and some say potential--in 1997, with both browser makers adding support for the technology and the International Organization for Standardization approving the second version of the specification.

Since then, VRML has acquired an unfortunate distinction as one of the Web's most hyped technologies not to live up to its promise. The VRML Consortium kept the flame burning for the specification but last year decided to tone down its association with VRML with the name change to the W3D Consortium and a new approach to the technology.

This year saw a burst in membership for the rechristened consortium, as about 20 members joined to boost the roster to about 65 companies.

X3D resembles VRML, except that it is built in the flexible Extensible Markup Language (XML), a World Wide Web Consortium (W3C) recommendation for labeling and structuring data and for designing task- or industry-specific markup languages, such as MathML for mathematical Web functions.

X3D has scored significant industry support from players including Sun and Microsoft, as well as a pledge of support from the W3C for the new specification, which is expected soon.

In addition to Sun's Java3D, another implementation of X3D was launched this year by W3D Consortium member Shout3D.

The W3D Consortium faced a disappointing setback after a hoped-for open source contribution of a VRML browser and tools was withdrawn.

The VRML tools, originally developed by Cosmo Software and Intervista, were acquired along with those companies by Platinum Technologies. When Platinum hit the skids this year, the company in February signed a letter of intent to put its VRML technologies into open source development in conjunction with the W3D Consortium.

In open source development projects, the underlying source code for software is made publicly available, usually for free and licensed use. Volunteer developers generally post their fixes and innovations for use by the Net at large. A grassroots example of an open source project is the Apache Web server; a corporate-sponsored example is the Mozilla.org development of Netscape's Communicator Web browser.

But before Platinum could hand over the keys to its 3D software, Computer Associates acquired the company, throwing the open source contribution into limbo before an agreement was finalized.

The W3D Consortium is in the midst of trying to re-negotiate an open source hand over with Computer Associates, according to the group's president, Neil Trevett. Computer Associates did not return calls.

3D's open source limbo was a key disappointment of 1999 for some 3D technologists.

"From my point of view, 1999 was not the best year for 3D," said VRML co-creator Parisi. "Open source would have given a real kick in the pants to the industry had Computer Associates followed through on it. That was a real setback."

Although he acknowledged that VRML has failed to gain mass-market acceptance, Parisi noted that the 1997 VRML 2 specification lived on in various corners of the Internet, facilitating collaborative design and being used as an interchange format for graphics applications.

"It's the only format out there that all the tools understand," Parisi said. "It's alive, but [it's] off in these strange little pockets of activity."

Parisi expressed optimism for VRML's successor, X3D, but warned that it faced the same market hurdles of its predecessor.

"Even if the technological approach is totally solid, the question is still what they asked us for five years: Who the hell's using it? Show us the customers."

Parisi singled out Sun's Java3D as an example of a Web 3D technology with a potentially sound business model.

Meanwhile, the W3D Consortium and its members are following the early VRML philosophy that an open standard is the best road to a successful 3D technology for the Web.

The W3D's Trevett has called open standards "the foundation of the Web itself." But WildTangent CEO St. John argued that a technology without a market was too immature to benefit from an open standard.

"Are you creating an open standard because it's useful and valuable or for the sake of having an open standard?" St. John asked rhetorically. "Nobody has defined a need or value for 3D on the Web."

To back up his point, St. John gave as examples the two companies that have created de facto or near standards for multimedia on the Web: streaming media market leader RealNetworks and animation software maker Macromedia.

"The people who made Real and Macromedia products had very comprehensive visions for a certain domain of multimedia content delivery," St. John said. "Because they were able to control content and development, they could make it work."