Often characterized by internal squabbling, Microsoft's various groups working on 3D graphics technology for the PC and the Web have delivered more confusion and delay than product to market in recent years. And the fates of their various efforts are often so nebulous that Microsoft's employees, partners, and customers cannot agree whether they are living, dead, or, in the words of one insider, left "just twitching."
Hanging in the balance is control over technologies that proponents have long promised would revolutionize computing online and off. To date, however, they have yielded little to the average consumer. The challenge for Microsoft and its competitors in the computer and Web graphics market is to maintain the technological upper hand in anticipation of widespread consumer demand that has yet to materialize.
Within Microsoft's graphics world there are three technologies in various states of limbo. They are Fahrenheit, a project to fuse parts of Microsoft's and SGI's graphics programming interfaces; Chromeffects, a set of technologies to bring graphics common on personal computers to the Web; and Talisman, a graphics chip architecture.
All three of these technologies, like many in the same category that have never seen the light of day, have emerged and then foundered in a swift, quixotic market for computer and Web graphics. In addition to facing elusive demand, Microsoft has had to contend with smaller firms that have been able to maneuver around Microsoft in getting their products to market at cheaper prices and with lighter bandwidth demands.
Microsoft has not been alone in its struggle with multimedia graphics technology; just yesterday, Intel said it is abandoning its graphics chip business.
Microsoft's most successful effort in graphics is its DirectX technology, created to entice game developers into developing for Windows instead of SGI's OpenGL application programming interface (API).
Looking ahead to a world defined by the Web and open standards, SGI began developing the Virtual Reality Modeling Language, or VRML, which became an official standard issued by the International Organization for Standardization. SGI based its Cosmo player on VRML. Cosmo was later sold to Platinum Technology, which tried to resuscitate it with an open source project. Computer Associates subsequently acquired Platinum.
Throughout this turmoil, VRML has entered a limbo of its own, failing to achieve widespread implementation. But three years ago, when 3D on the Web seemed to be the next hot area, Microsoft saw the need to counter it and other Web-based threats. Enter Chromeffects, a set of high-level programming interfaces and simple coding tags that would have made it easier for Web developers to access DirectX APIs.
Microsoft previewed Chromeffects in July 1998. In November, the company sent the project back to the drawing board in response to negative developer feedback and in December said parts of the technology would come out in its Internet Explorer Web browser, with the crucial set of XML coding tags to follow after shipment of the Windows 2000 operating system.
Microsoft today maintains that Chromeffects is alive and well within IE and other products.
"People get caught up on this standalone deliverable called Chromeffects," said Shawn Sanford, group product manager for consumer Windows at Microsoft. "What we found is that maybe a single standalone deliverable was not the best way to get the technology to market. To say that the technologies are dead is false."
But Microsoft's partners, developers, and original members of the Chromeffects team share the view that however Microsoft salvaged its parts, Chromeffects failed.
"The people who were creating that technology are gone," said one source close to Microsoft. "They shipped some pieces in IE 5, but there are no armies of evangelists going out and promoting the technology, and Microsoft isn't teaching developers how to use it. That means it's dead."
Analysts, partners, and developers said Chromeffects required too much processing firepower to be practical for most consumers. Developers objected to its lack of standards support and said it was generally cumbersome. And many speculate that Microsoft pulled it in part to evade further scrutiny from federal antitrust investigators for pushing a Web-based product that would have worked only with Windows.
Others said that in abandoning Chromeffects, Microsoft was paying attention to its bottom line.
"Chromeffects ended for many reasons," said a senior executive of one of Microsoft's partner companies. "None of the [computer manufacturers] wanted to pay for it, especially with all the interest and growth in low-end machines. Microsoft makes just as much money on a cheap computer as it does on a really expensive one. Chromeffects didn't drive that."
Sanford refuted all the above arguments, saying Microsoft acted only in the best interests of its customers and partners.
3D Web graphics: Who's buying?
Another hurdle facing multimedia Web graphics is the dearth of consumer interest in them. Taking Yahoo as their model, many Web designers and particularly businesses aim primarily for light pages that download fast. For this reason, Macromedia's lightweight Flash animation system has made significant inroads in the market.
Even if multimedia Web graphics seems like a nonstarter for now, it would be uncharacteristic of Microsoft to sit by while competitors define the technology. In addition to Macromedia, whose product typically is used for simple animations, Sun Microsystems is implementing its cross-platform Java3D technology, and the Web3D Consortium is pushing X3D, a successor to VRML.
Sanford maintained that the market for multimedia graphics remains "huge" both on and off the Web, citing markets for computer-aided design (CAD), games, and graphics cards, and stressed that Microsoft continues its work on the Chromeffects technology.
"Microsoft is now an active participant in the W3D Consortium working to define X3D," said Neil Trevett, president of the consortium. "Our current draft takes lots of good stuff from Chromeffects, particularly the way it integrates with XML. Microsoft has been very open in sharing lessons good and bad from its experience with Chromeffects."
While Chromeffects quietly fades from view, graphics analysts have focused their attention on Microsoft's Fahrenheit initiative to build a standard that blends aspects of DirectX and SGI's OpenGL programming interface.
Fahrenheit emerged as part of a 1997 peace treaty between Microsoft and SGI ending their "API wars," in which the two companies fought for dominance of their graphics programming interfaces. In addition to collaborating on the graphics standard, the two companies allied more closely as SGI embraced Microsoft's Windows NT operating system.
Three years later, Microsoft no longer has to contend with SGI as a significant threat, and SGI is moving away from NT toward the Linux platform. Coupled with SGI's recent layoffs and restructuring, this changed landscape has left many wondering how much the companies have at stake in bringing Fahrenheit to market.
"Fahrenheit is quietly moldering away at Microsoft," said one source close to Microsoft's graphics efforts. "Fahrenheit is going away quietly, the same way Chromeffects did, the same way Talisman did."
After SGI announced significant layoffs and restructuring this month, some speculated that Fahrenheit would die on the vine while SGI refocused its attention on profitability.
SGI declined to say how many of its Fahrenheit engineers were included in the layoffs, which number between 1,000 and 1,500 out of 9,000 employees. The company transferred about 30 engineers to Nvidia, with which it formed an alliance, to work on a graphics chip.
A spokesman denied that SGI was withdrawing its support for Fahrenheit. He speculated that some observers may have confused its withdrawal from one chip design project, which it transferred to Nvidia under an alliance between the two firms, with its commitment to graphics technology in general. That commitment remains strong, he said.
"We are not getting out the graphics business," said the spokesman, who asked not to be identified. "We remain very dedicated to it. It's part of our heritage."
Microsoft also reiterated its commitment to Fahrenheit following a June 30 statement by Microsoft chairman Bill Gates that Fahrenheit work is moving ahead and that "a year from now, we should see those results."
But it is unclear how much of a cooperative effort the project remains.
"We're going to continue to move forward to continue to work to enhance our graphics technology to bring those benefits to our customers," said Microsoft's Sanford when asked to characterize the collaboration with SGI. "We are committed to moving forward with our software efforts."
Talisman was the code name for a graphics chip architecture that Microsoft introduced in 1996 that was supposed to have offered 2D and 3D graphics, Windows acceleration, full-resolution MPEG-2 playback, videoconferencing, sound, and modem capabilities on a chip costing less than $300.
As with Chromeffects, Talisman never received an official burial but was salvaged for parts. Some of its technologies were incorporated in DirectX APIs, and others into various chips, according to Microsoft.
Talisman died because the market shifted underneath it, according to analysts and people involved with the project. Microsoft proposed it to overcome bandwidth limitations and the need for memory in graphics display chips. But in the time it took Microsoft to develop the technology, memory prices evaporated, bandwidth increased, and chips improved independent of Talisman.
"The problems it was meant so solve went away," said Jon Peddie of research firm Jon Peddie and Associates. "Also, the Talisman solution was itself intrinsically expensive, so it became more expensive than it was worth."
Microsoft does not deny that the graphics market has moved in unexpected directions. And the practice of delivering technologies in bits and pieces reflects the realities and opportunities inherent in the software business, Sanford said.
"In the software industry, plans change," Sanford said. "The market evolves very rapidly, and what I think may be the best way to deliver technology today may not be the best way in three days. One of the really neat things about the software industry is your ability to adapt to the changing conditions. You can take market feedback and deliver what best serves your partners and customers."
Ultimately it may wind up being a matter of interpretation whether Microsoft is killing its graphics projects, putting them on the back burner, or simply devising alternate means of delivery.
"The Talisman project faded away," said one person who was close to the project. "It was a crazy idea, and the partners bailed out. When Microsoft kills something, it never says, 'We're dumping the project.' It just fades away."