Acacia Research Group

CONFERENCE REPORT: 3D On the Web & Beyond
By David Duberman

Whither Web 3D? The oft-asked question was posed most recently at 3D On The Web & Beyond, a conference hosted last Wednesday and Thursday in San Jose, Calif. by Texas-based Acacia Research Group (ARG). The small group of conferees didn't arrive at any conclusive answers, but lots of information was conveyed, and presumably most came away somewhat enlightened as to where things stand and which way(s) they could go. This was the second year of the conference, which distinguishes itself from similarly named technology-oriented events at Siggraph by its emphasis on business. And one thing the folks in San Jose agreed on was that business is down. VRML co-creator Tony Parisi ventured that this is a good thing because it forces people to get serious about what they're doing and stop pursuing "crackpot" ideas. The phrase "crossing the chasm" was heard often at the conference; of course, how and when we'll get to the other side remains to be seen.

The conference took the form of a series of panel discussions alternating with short presentations from representatives of various firms and organizations. As usual at such events, most everyone who spoke had an agenda or something to sell, so the trick for observers was to separate the hype from the information. Among the participants in the first panel ("Where Do We Stand?"), other than Parisi, was Neil Trevett, who wore three hats at the conference: VP of 3Dlabs and president of both the Web3D Consortium, which is creating X3D, the putative heir to the VRML throne, and the Khronos Group, which is working on a version of OpenGL for mobile devices. Trevett stated that niche applications will serve the primary role in bringing Web 3D into the mainstream, to which Parisi responded with his "chasm theory," according to which corporate customers are willing to pay a lot of money for Web 3D, so one can earn a living by working with one at a time. In this, Parisi should know whereof he speaks: He currently heads Media Machines, a developer of "high-performance, standards-based 3D and integrated rich media core technologies."

Of course, when you talk about the current (and probably future) state of Web 3D, to a large extent you're talking games. The panelists were asked whether 3D games online could lead to other Web-based 3D applications, to which Miriam Geller, the Macromedia product manager in charge of Director and Shockwave, responded that the most popular Web games are in 2D, and are fast, easy, familiar, and let the user in and out quickly. That was when Alex St. John first spoke up. Every good conference has at least one larger-than-life participant who serves as a class clown/lightning rod, and the erstwhile games evangelist at Microsoft played the role to the hilt at this one. At one point, when speaking from the audience, St. John was offered a microphone, to which several people responded simultaneously that he didn't need one.

As founder and principal of WildTangent, which sells technology and services for online gaming, St. John was the primary gaming proponent at the conference. Much of what he said was aimed at pushing his company, which got to be tedious after a while. But perhaps the most interesting observation he made, although somewhat off-topic, was that Intel's AGP bus doesn't adequately serve its intended purpose of quickly moving data between the CPU and the GPU; the equivalent bus in the Xbox is much faster, according to St. John. Thus, he claims, high-end consumer-level PC graphics cards are far more powerful (i.e., expensive) than they'd need to be if the AGP bus were more efficient. St. John's response to Geller's statement was that the typical consumer can't tell the difference between 2D and 3D, so the added bandwidth and overhead required by 3D games offers diminished returns. This, despite the fact that WildTangent's technology is fully 3D-based. Moderator Samantha Staples said that the ideal non-game application is 3D assembly instructions for products like Ikea furniture, to which St. John responded that video-friendly broadband is the enemy of 3D for that purpose, because video is easier to produce and distribute. Geller had the last word, saying that despite this, 3D works better for online instruction because it's more flexible.

The next panel was Repurposing And Building 3D Content For The Web, which Jim Giebutowski of SolidWorks kicked off by stating that the key issue is not technology or content-creation applications, but getting the content created. Many companies trying to sell on the Web don't have 3D aspects, or even 2D assents. If they do have 3D assets, those might contain proprietary information, such as the engineering data that can be embedded in SolidWorks models. Such information would need to be stripped out before content could be offered to the public online. Panelist Rick Benoit of Intel added that the most important phase is gathering end-user requirements and determining whether a new platform is needed or an existing one, such as X3D, will suffice. Giebutowski asserted that the major benefit of 3D is to be able to do things in an ad hoc way, at your own pace. Of course, this is more a consequence of interactivity, which needs to be implemented properly, but 3D does play an important role.

After this, Neil Trevett provided an update on Web 3D Consortium activities, and predicted that the first version of the X3D specification would be finished by the end of 2003. He also mentioned that the organization's next consortium would take place in France on March 9-12, 2003.

Next up was a panel on the future of Web 3D applications. Ken Deaton of 3DNA showed his company's 3D user interface for Windows, which resembled a large building interior that seemed to require far too much navigation to achieve the same results you get with minimal pointing and clicking in standard Windows. Deaton's premise is that productivity has to do with working in a space that you're comfortable with, and he claims users will get into customizing their spaces and will build communities based on trading assets. Don Brutzman, who, as a faculty member of the Naval Postgraduate School, represented the U.S. government at the conference, offered an intriguing statistic: The 2001 federal information-technology budget was $44 billion, of which Brutzman estimates 15-20 percent involved 3D graphics.

The last panel of the first day was Successfully Selling 3D For The Web. Most of the panelists were from companies like Viewpoint and Cycore that sell high-priced technology and services; Joe Wilson distinguished his employer by pointing out that it sells an authoring tool, Anark Studio, which costs $600 (competitive upgrade) and for which there are no license fees. The consensus on this panel was that, far from trying to market 3D as an all-in-one solution to enterprise customers, the wise course was to sell 3D as a piece of the "puzzle," of which other components include rich media, interactivity, etc. Here, as elsewhere, there was much talk of standards: the need for proprietary players posed by many of the solutions offered by panelists' companies seems to stand in the way of Web 3D's success, but is necessary so that companies can innovate. Neil Trevett, wearing his Web 3D Consortium hat, acknowledged this need, but said that there's also a place for the portable, lightweight X3D.

As Thursday's opening keynote speaker, Trevett informed the audience about the status of the Khronos Group's efforts to produce a reference implementation of OpenGL ES (embedded subset). The group settled upon OpenGL because it's a widely available, open, low-level spec that's well proven, successfully implemented, and free of royalties. The footprint can be greatly reduced by eliminating features not necessary for mobile devices such as workstation functions and volumetric (3D) textures. Trevett issued a call for participation in defining the spec: go to http://www.khronos.org/ for more info.

Most of Thursday's panels were about the "Beyond" part of the conference's title; namely, getting 3D onto mobile devices. Among the more impressive demonstrations were of 3D performing reasonably well on cell phones. Mike Pell of Microsoft showed his "Space Needle," a combination cell phone and Pocket PC, while NetImmerse's John Austin showed an application that lets users walk through a real estate development. Austin readily admitted, though, "it's really going to be all about games." An intriguing idea about games on mobile devices is that they can incorporate the player's location; how that will play out should be interesting. Alex St. John, who candidly states, "I'm just interested in opportunities that make me a billionaire," said that 3D for mobile devices will take a long time to mature (two-three years); currently, the screens are too small, the buttons badly configured, and video is equally compelling for many apps. To resolve this, Intel is working on a design spec, to be built by others, of a clamshell phone/gaming device with a button layout specifically designed for gaming.

The last panel of the day, and of the show, returned to our roots with a look at 3D tools. Among the participants were representatives from Alias|Wavefront (Maya), Discreet (3ds max), Newtek (LightWave 3D), Alias|Curious Labs (Poser), Maxon (Cinema 4D), and Realviz, which makes a tool for generating 3D models from photos. Probably due to the limited time available, the panel amounted to little more than cheering sessions for the participants' products. Geoff Foulds of A|W offered a look at digital-content-creation survivors, who work efficiently, minimize the number of tools in the pipeline, and maximize the range of output quality from 2D to film-quality, real-time, CG effects. He predicted the demands for quality will increase, but also said the future is "simple."

Arnie Cachelin of Newtek predicted the return of the 3D professional, and said that Newtek will provide exporter code in a free, public SDK, while keeping Web 3D development on the back burner. Discreet's Dave Campbell in his marketing mode pointed out the success of his company's various products; the company has a user base of over 300,000, and if you add the 200,000 downloads of the free gmax tool for gaming add-ons, over half a million. Campbell also singled out Anark Studio as mixed media that doesn't look like the Web (i.e., IE). Maxon's Paul Babb stated that his company aims at "entry-level" users, and succeeds because of its dedication to customer service. And Steve Cooper of Curious Labs pointed to his 120,000 users, touting Poser's ease of use and plug-ins that let it work with other 3D apps. And that was that.

Despite the overabundance of hype, the conference was an interesting, educational experience; an opportunity to learn what's coming down the pipe and grab the lapels of major players in the diminished Web 3D industry. It also served to introduce Acacia to those of us less familiar with the group; they brought along copies of their reports on 3D content creation and mobile devices for attendees to browse.

http://www.acaciarg.com