Digerati - Chapter 20

Digerati - Chapter 20

John McCrea [10.1.96]

Chapter 20

THE FORCE

John McCrea

THE WEBMASTER (Kip Parent): John McCrea has been a tremendous force at Silicon Graphics. He was able to rally the engineering and marketing people in the company around the Web, and he managed to build the WebFORCE product line into a rapidly growing business within the company. It is a tremendous success. John is definitely someone who sees the big picture and has a vision. He rallies people around him and makes things happen. 

John McCrea is the manager of Cosmo, Silicon Graphics's next-generation Web software product line.


Networks and computers have been around for a long time, and network computers have been connecting people around the planet for a long time. For Silicon Graphics's John McCrea, the connectedness that is occurring and accelerating right now is interesting. John is one of the people leading the drive toward a new medium. He sees the Web as the key enabler to bring about 3-D hypermedia: virtual reality modeling language, or VRML. He's not the typical Silicon Valley guy. I know John because through his efforts, Silicon Graphics has provided technology support to Content.Com, Inc., the Internet publishing company I have formed with David Bunnell. Part of his job is to seek out the right start-ups (HotWired, for example) and provide Silicon Graphics technology as a way of getting the word out about his products. Unlike many of his colleagues, he lives in San Francisco, and his interests are oriented more toward the arts than toward engineering. It's refreshing to talk to him.

John entered the computer business late, joining Silicon Graphics in 1993 right out of Stanford's MBA program. In 1980 he went to MIT with the notion of becoming a physicist and fairly quickly realized that there were lots of other people who would be better physicists and that it was not an exciting time to be in physics. His second passion was the arts, and strangely enough, it was possible to pursue creative writing and other art forms at MIT, a school known for engineering and science."The '80s were the lost decade for me," he says. "I was doing everything from moving to Ireland to work on a novel to doing fund-raising for a school in Oregon. In '91 I had a 'born-again' capitalist experience, went to Stanford business school, and fell in love with Silicon Valley. I worked a summer at Tandem, then ended up at Silicon Graphics. So I am not a computer industry insider, by any means."

In early '95 John thrust Silicon Graphics into the Internet market with the WebFORCE server product line. Today he is manager of Cosmo, their next-generation Web software product line. "Who knows where that will lead to." he says. "I was the product manager for the Indy workstation, and the Web thing came on really strong. It was obvious that we had an opportunity, and I went forward and defined what we would do and pulled together the team that would build the first Web authoring tools and the first Web system. It has grown into a multi-hundred-million dollar business for us, in both servers and authoring workstations. I lead much of the business, and the marketing effort behind it, and I work with the engineering teams that are building the next generation of products."

THE FORCE (John McCrea): At a technical level, VRML is a file format. At a higher level, it is a way of doing 3-D graphics over networks. 3-D graphics are coming to a Web where people have become comfortable with a page-centric view. But VRML will take the page-centric view and pop it into another dimension, with the potential to make the experience more like the physical world. This fundamental shift brings a sense of place to something that has absolutely evaporated the notion of space. While evaporation of space is a powerful concept, much of the information suddenly loses context. Introducing the notion of space to the Net via VRML has the potential to make it more compelling and appealing to a larger audience.

Increasingly you will see 3-D objects that you can move around and look at from different angles right inside your Web window. Your Web page will have text and pictures and 3-D in a little window. This is the intermediate step to a more major shift that happens when, instead of encountering 3-D as if it were just another type of media, like a picture or a movie clip, what you see in the window is a 3-D world. This world is like the outside world in that you can choose where to go and steer yourself through it, flying or walking or whatever. You get to where you want to go in the same clicking fashion as with hypertext, but now it's hypermedia. I click on a door and stuff comes to me. I click on a window and stuff comes to me. That stuff could be more 3-D or a video clip or a picture or text. It is a major shift in your experience of that information set. It is not better than text. It's different.

Rikk Carey, one of the original leaders of the VRML effort at Silicon Graphics, contrasts the Web with a library. The differences are significant. On the Web you can't go to a section of the library and see all the books related to a topic, or see how big they are, or get a sense of the style of the book from its cover, or what it looks like inside. You can't even go inside the library; you have to stand outside. By typing an address or hitting a link, you shout up to a window and ask for a particular page. Mysteries happen behind the scenes, and moments later the page you requested is held up to the window. You never see the book and you never see the chapter. You see only the page.

The Web today, a world organized by bookmarks, is a very different model from what we human beings, who live in a 3-D world and have a spatially oriented memory, are used to. When you recognize, for example, a picture of a house in a field, you immediately know something about that house and field. Suspending your disbelief, you think of the real world. When you are enter a 3-D environment under your own control and can move through it as a virtual world, you have to get a certain frame rate up, considered to be a minimum of twelve frames per second for reasonably smooth motion. At that point, your mind is no longer interpreting a still image and figuring out its worldliness. When you're moving through something, you're experiencing it, and a powerful cognitive shift occurs.

At Silicon Graphics, we searched for how we might be able to make VRML happen. we were working for years toward what would eventually be called VRML. That effort was originally called "Inventor." Inventor was the first object-oriented toolkit for developing 3-D applications. It was open and cross-platform. When Mark Pesce and Tony Parisi were talking up the vision for VRML, the search for the technology brought forward Gavin Bell, a member of the Inventor team at Silicon Graphics. He drafted a proposal based on Inventor that became known as VRML 1.0. The spec came out in early 1995 and, probably most important, was adopted by a small group of people who felt passionate about it. At Silicon Graphics a small team broke off and started working on the VRML products. That effort gave birth to the VRML 1.0 browser.

Silicon Graphics basically took the lead role in starting a new industry around VRML. The specification, the first parser, and another piece of enabling technology were thrown out over the firewall into the public domain, and about seventeen different companies started making VRML browsers because it was relatively easy to do. Now at least a half-dozen companies have VRML products based on that technology. Microsoft was able to get one by licensing it from a small company.

Where do we go from here? We now have worlds that you look at and fly through, and are inhabited by creatures. One of the first was fish swimming, with sound that is spatial and 3-D. Moving around, you feel you are in the space. Creatures come up to you, and if you click on them, they do something different. What happens when I, as a viewer of this content, can begin to build this myself, when I can take different, almost robotlike parts and put them together? I can start to have things that represent me in this world. Or I can interact with a representation of someone else in the world. It really is a shift, a blurring the line between application and content.

The whole notion of live content is like the notion of artificial life. It's not something that people are thinking about right now, but wait until you start to exist in these worlds. You'll be walking down the street of a city that doesn't exist, and walking into stores that don't exist, but interacting with people who do exist, represented to you in a unique way on a computer screen and with agents that are programmed animated characters. We are at the beginning of trying to figure out what it will mean. What happens to the robot avatars when no one is looking?

Microsoft is a company to be taken seriously. Its fortune has been built on proprietary formats and closed APIs, but the Internet is forcing it to behave in a different way. A number of battles will be fought, and Microsoft has stated an intention to go open, a complete reversal of strategy. The evolution of VRML, from 1.0 to 2.0, was a defining moment in the new era of competition around open standards. Microsoft proposed something called Active VRML, which is in no way technically related to VRML. It was a complete reset of how one might do 3-D. Of course, innovation is great, so the more the merrier, but the way Microsoft presented Active VRML raised questions and concerns within the online community. They haven't figured it out, but they want to throw something out there and slow the process down and confuse the market. But the community of interest around VRML rallied and joined battle for VRML 2.0, which was initiated by Silicon Graphics proposal called Moving Worlds. The battle between Moving Worlds and Active VRML is now playing itself. Moving Worlds is fast becoming the standard with support from Netscape and Apple. VRML 2.0 products such as Silicon Graphic's Cosmo Player are now coming to market. VRML is an open file format, and because 3-D is complex file format actually has a lot of implications for what you need to make an authoring system and a viewer. At the end of the day, VRML is a file format. Silicon Graphics invented the file format, essentially extracting elements from Inventor.

The rivalry between Java and VRML got a lot of attention. The reality is that they are not competitive, but are perfectly complementary. VRML is a file format for doing 3-D, and Java is a programming language. Early in the rise of hype and interest around Java and VRML, both sides, Silicon Graphics and Sun, started thinking about how they could be complementary. Both companies are cooperating or mutually endorsing each other's technologies. In fact Silicon Graphics is working with Sun to bring 3-D and multimedia to Java, based on the work with VRML. We're bringing to market Cosmo 3-D, which is a VRML 2.0­compliant API for the Internet. Cosmo will allow developers to create platform-independent 3-D applications in Java or in more traditional programming languages like C++.

People are making interesting analogies about the Web. One is that the Web is becoming the world's largest disk drive. I like that, except it doesn't paint a very exciting picture of the content. The analogy of the world's largest library doesn't excite most people because they find libraries fairly boring. Another analogy is that the Web is more like a CD-ROM than a disk drive. But with CD-ROMs, you can't change the information. The Web is more like a CD-ROM with immediately updatable information through databases. At Silicon Graphics, we're trying to take the Web from a static, page-centric, download-then-view model and make it dynamic and experiential, where the bandwidth will eventually make it totally live, with streaming. A lot happens when you go from page-centric to worldlike with the combination of 3-D graphics and audio, and the interactivity that Java brings. These technological changes are happening so quickly that one of the great challenges is to match a marketing strategy with product development.

Because the Web is evolving so rapidly, our strategy has to morph at a minimum every three months and in some ways on a weekly basis. A year ago Silicon Graphics focused like a laser beam on the Web because we saw the Web as uniquely visual and engaging‹different from the Internet that had come before it. Now we are bringing our core technologies in 3-D graphics and streaming media to the emerging open platform of the Internet. We focused on the problems people were having in implementing homepages or sites on the Internet. When you view the World Wide Web as a medium, you realize that there are only a couple of ways to make it economically viable: a pay-per-view model or advertising. In the case of a corporate site, the content needs to be compelling enough that people want to come to what is essentially a corporate message center.

Creating engaging content that generates interest is where you get economic value. When you make content that is media-rich, you need the tools to build that content, and you need servers to serve that content. The more media-rich it is, the more traffic it generates, the harder a job the Web server has. Silicon Graphics's value proposition for that whole space is the ability to make the best possible content, with servers designed to get that content to a large number of people efficiently and reliably. Because databases are now involved, it is not enough just to get up a homepage. Now you need to have a Web site where you're doing real business. When you're doing real business on the Web, you want to capture information, you want to customize the content, and you want to do transactions.

Whether you can do a hundred thousand, one million, or ten million hits a day on the server is an important consideration, but database performance and reliability are also concerns. Another concern is the applicability of Web technology to the enterprise, or to business processes that have nothing to do with sending out a corporate message or generating advertising dollars. The Web can become a universal front end because the Web client is really the world's first universal interface to information and applications.

The growth of this part of our business is tied to an interesting synergy between Silicon Graphics and our customers. First, the Web browser spread like wildfire onto every desktop at our company. Then we created the first set of tools for easily building Web content, and they spread like wildfire. Now we have a situation at Silicon Graphics where outside the firewall there is a handful of Web servers; and behind the firewall there are more than two thousand. So the whole company is being run with the Web. Existing customers and potential customers want to know how we do it. Innumerable executives from around the world have traveled to Mountain View, California, to see how a true information-age company is run. The powerful notion behind this is that operating system religion is irrelevant. You can have whatever operating system you want. You can build Web infrastructure with much of your existing infrastructure.You don't have to throw out your mainframes or your Unix database servers or your Macintosh desktops. It's a wildly exciting time. People looking at this Web stuff for the first time find it a little intimidating because so many different technologies have been brought together. In reality, putting the Web to work for their business can be implemented very easily. A number of our customers have borrowed our systems, for example, a workstation with all the Web-authoring tools and site-management tools, for a weekend and in forty-eight hours they have implemented the beginnings of an intranet.

Creating the best possible content outside the firewall is the issue. The IS department‹those unfortunate souls who have to build the information infrastructure for a company and are usually blamed when there are problems and rarely thanked when things run well‹doesn't care about the best possible content. They care about needing to train, retrain, and hire IS peoples. When we show them that we have two thousand servers up internally, and two-hundred thousand URLs, all of which are managed by five people, they are amazed. These five people can do it because the Web has an unusual balance of centralized control and decentralized empowerment. Everyone can create their own content, and the IS organization, if it wants to, can control the framework in which that content gets deployed.

I first heard the word content used in the way that I think about it now in 1993, when the notion of multimedia was rippling through the Valley. Content is one of many interesting words that are now commonplace, at least for those of us in this industry. When we were launching WebFORCE, Silicon Graphics's product line for the Web, I thought about how to describe a professional Web authoring and serving system to somebody totally removed from Silicon Valley. Think of all the words that need definition. Web. Authoring. What is authoring? Even though Content and authoring are old words, the way they're used now seems to suggest something different. As technology changes, we need to develop new ways to express ourselves. Otherwise, we only see ourselves through a rear-view mirror. I've been struggling to find the tools, the vocabulary, the instrument that could express exactly what I'm thinking and feeling about the changes in communication we're currently experiencing. Somehow, I don't think I'll find it in the written word.

THE SEER (David Bunnell): John is not really "one of us." He is a well-rounded, cultured, artistic, educated person who just happens to be at Silicon Graphics. He would be successful anywhere.

THE CONSERVATIVE (David Gelernter): McCrea is right at the center of one of the hottest technologies‹Web servers‹of modern times. The machines he markets are the foundation of the Internet; the towers that support the bridge.


Back to Contents

Excerpted from Digerati: Encounters with the Cyber Elite by John Brockman (HardWired Books, 1996) . Copyright © 1996 by John Brockman. All rights reserved.