Overdrive, page 11
One of these random efforts was an Internet-related project under the direction of James Allard, and known by the acronym TCP/IP, for Transmission Control Protocol/Internet Protocol. For years, Bill Joy of Sun Microsystems had been putting TCP/IP into his company’s UNIX-based computers. Allard’s job was to develop TCP/IP for Microsoft’s Windows platform. His orders had come directly from Steve Ballmer, who had discovered during an out-of-town sales trip that some of Microsoft’s best Fortune 500 customers were better connected to the Internet than was Microsoft. The nontechnical Ballmer didn’t know what TCP/IP was, but he knew he wanted Microsoft to have it.
Allard, who would become one of the first within Microsoft to beat the drum for the Internet, said the message he got from Ballmer about TCP/IP was: “I don’t know what it is. I don’t want to know what it is. [But] my customers are screaming about it. Make the pain go away.” So in early 1993, Allard took charge of the development of Microsoft’s first Internet server, which would connect Microsoft to the Internet and distribute test copies of its TCP/IP to customers.
In addition to Allard’s project, Microsoft had a business arrangement with CompuServe, which sponsored an on-line forum where customers could get technical support for Microsoft products, for which Microsoft received about $1 million a year in fees. And Microsoft’s Exchange group, which was developing an e-mail and groupware product, was talking about an on-line deal with AT&T. Even the folks working on what would become BOB, the ill-fated project to develop a social interface using cartoon characters to help make applications easier to use, were talking with MCI and others about an e-mail server. (That group of managers included Gates’s wife- to-be, Melinda French.) And, of course, there was Jeff Lill’s random on-line effort, which would become the nucleus of Siegelman’s project once it got the okay front Gates.
“Russ was really doing two levels of things for Bill,” said Lill. “He was reporting on all these random efforts that were going on within Microsoft, and he was going outside the company and meeting with CompuServe and others to get a feel for what was out there before going back to Bill with his recommendations. Once Russ determined that Microsoft would not be able to buy off the big services like America Online, it was clear we would have to build our own.”
In early 1993, AOL had only about 350,000 subscribers, which made it less than a third the size of rivals Prodigy and CompuServe. But AOL boss Steve Case had decided to put the pedal to the metal to overtake his competitors. AOL began flooding the mail with free sign-up software specially designed for Microsoft’s Windows, and it went to work signing up content providers, including magazines and TV networks. New material was added almost weekly.
AOL was not the only on-line company Microsoft tried to buy as a way to jump-start its entry into the market. It also was interested in a buyout of Sierra On-Line, which was headquartered in nearby Bellevue. But that deal didn’t pan out either.
“If we could have bought something, we would have,” said Lill. “The AOL thing was for a significant amount of money. I don’t know how much, but it was significant. The initial idea was that we would just go out and buy a company and get into the service that way.”
But even though it had to start from scratch, Microsoft believed it would gain the advantage later by bundling its on-line service with Windows 95. The promise of millions of customers signing up for Microsoft’s service when they bought Windows 95 convinced Gates to give the green light. But it meant that Siegelman and his team would face a daunting schedule, because the service would have to be finished in time to ship with Windows 95, whose coming-out party was scheduled for June 1994. (That launch date would later slip to December, and finally to August 1995.)
Once Gates had given the okay, the next decision was where in the organization to put Siegelman’s development team. “Bill realized there was no real good place to put this,” said one Microsoft executive. “And there was no real division that wanted it, because it just didn’t fit with any of the existing divisions. So Bill just plunked it under Nathan. The Advanced Technology Group’s job was to provide a sort of nurturing for new stuff that wasn’t well defined enough to be in a real product unit with hard-core goals and things. So it was actually a fairly appropriate place to put this.”
After the project had been officially launched, it was time to come up with a code name, as for most Microsoft products under development. Windows 95, for example, was known in 1993 only as Chicago.
The story of how the Windows team chose to nickname Windows 95 Chicago began with a spur-of-the-moment decision in early 1992, when Jon Lazarus, then Microsoft’s vice president for strategic relations, was preparing to give a confidential briefing to software developers about an exciting new operating system that would be a successor to Windows NT. He decided to invoke an exotic locale—Cairo. The city theme immediately caught on, so when Microsoft began developing a successor to Windows 3.1 and needed a code name, a number of cities, foreign and domestic, were kicked around. “We wanted something between Seattle and Cairo in terms of functionality,” said Microsoft Vice President Brad Silverberg. “The less ambitious picked names closer to Seattle—like Spokane for a minor upgrade, all the way to London for something closer to Cairo.” Eventually, the name Cleveland was chosen. But that didn’t stick for long. “We knew Cleveland wouldn’t fly too well as a code name,” said Silverberg. A few hundred miles due west of Cleveland was that city of big shoulders, Chicago. That name stuck. Microsoft had good memories of Chicago, which had been the launch site of Windows 3.1. It was just coincidence, though a nice one, that some 375 miles south of Chicago was a city named Cairo. (Microsoft would continue the city theme when it needed a code name for the successor to Windows 95 and settled on Memphis because that was where Elvis lived. A minor upgrade was to be called Nashville, because it is near Memphis.)
Finding a code name for Microsoft’s on-line service wasn’t so complicated and had nothing to do with geography. Siegel- man’s team came up with Marvel by doing an on-line search of Prodigy’s thesaurus. The Marvel Comics name immediately fired up the team because of the prospect of naming servers and other components of the project after some of America’s superheroes, like Batman.
No Microsoft code name remains a secret for long, however; nor is it supposed to. Microsoft executives use project code names when giving confidential briefings to outsiders, and they eventually make their way into the public domain through media reports. Thus, Marvel became known outside of Microsoft early in its development. Consequently, the company received a letter from a lawyer representing Marvel Comics, who explained that the name was a trademark and Microsoft could not use it. But as far as Microsoft was concerned, the Marvel name was for internal use only and was not going to be the name of the on-line service when it came to market. The letter was ignored.
At the beginning, the Marvel team consisted of fewer than 10 people, including Siegelman and Lill. The team took over an office area in Building 8, where Gates himself worked. The team’s first mission was to come up with a plan for the kind of service they were going to build.
“We did some market research to get a feel for what the hell we were doing, for what the goal was,” said Lill. “We were going into this with the goal that it was going to be a computer user-oriented service, rather than a general-purpose service.
So, we were looking at our CompuServe traffic and saying ‘Hey, we’ve got a lot to offer a computer user. A lot of users are going to CompuServe to visit Microsoft support areas. We know something about that. So we, as Microsoft, have something to offer computer users.’ We were originally going to focus on being a very targeted product. And we did some very early prototypes that we showed some of these market research folks down in Los Angeles.”
Lill and Siegelman also flew out to Washington, D.C., to visit a tiny company that had developed some on-line technology that they thought might be useful in starting Microsoft’s service. The company was providing on-line support for software companies, with bulletin boards, e-mail, and a lot of the services that the Microsoft team thought they might want to include in their own, even though it was all UNIX-based.
Siegelman, who had a reputation around Microsoft as a first-rate negotiator, reached an agreement with the company for Microsoft to license its product, pending a technical review by Lill. But Lill recommended they pass on the license deal, even though he found it to be a very good product. He decided it just didn’t fit with the Marvel plan.
The initial design for Marvel looked a lot like Prodigy, which Lill soon realized was the wrong approach. “It was not very Windows-like,” he said. “We were trying to go for a very, very easy-to-use thing in the prototype. And just after that, I realized that we were reinventing the wheel. We would have had to create new user interfaces, and we were not a user interface group, and our overall goal was to ship something with Windows 95. And so we just didn’t have the time to be monkeying around and inventing things. And I had seen the BOB effort, and it was a black hole of changes and changes and changes and changes. So I basically told the team, ‘We are going to make this just be Windows.’ ”
The group was under the gun. Gates was supposed to be briefed on the project in late September. And he wanted more than a report; lill wanted a demonstration, proof that it could actually work if the company invested further time and resources. True, Gates had approved the project in May, but it would be the September meeting at which he would decide whether to go forward or scrap the effort. Thus the team had to be able to convince Gates at that meeting that they could do this. But could they?
A number of technical problems had to be resolved before the fateful meeting, the most important of which was to decide the underlying infrastructure for using an X-25 network. This was an older technology, but very similar to the TCP/IP that Allard was working on. AOL, Prodigy, and CompuServe all used the X-25 network for their on-line services.
The team by now was also thinking about content. What would the service offer customers? Because it was being designed primarily for computer users, the initial intent was to follow the model of CompuServe and make various computer- oriented publications available. The service would also, of course, need a traditional e-mail function and bulletin boards. The plan was to use Exchange for both these capabilities. “All that was pretty much the core technology of what we were looking at to deliver with the initial service,” said Lill. “We figured that since it was mostly for computer-oriented users, they’d be a little bit more forgiving.”
It was Lill who gave the crucial briefing to Gates at the end of September. The team managed to get a simple demo functioning on an X-25 network. They even had some image-compression technology working, and downloaded ap. image using a 9600 baud modem. “We were really just trying to show Bill that we had the technology figured out,” said Lill. “We had the scary modem figured out and the scary network figured out, and we actually had a server running and making this thing work.”
As he always does at such briefings, Gates asked some tough technical questions. But overall, the presentation went well. Not once did Gates scream at anyone—as he had been known to do at these briefings—or throw something, or shout his familiar phrase, “This is the stupidest thing I’ve ever heard of!”
Nevertheless, Gates was highly skeptical that the team would be able to meet the June 1994 shipping date for Windows 95. He didn’t think it could be done. He also had some nontechnical questions. “How much is all this going to cost?” he wanted to know. “Can we really make any money?”
Lill did not have ready answers for those questions, but Gates gave his approval anyway and authorized the hiring of some 30 more people for Marvel.
Which raised the question of where to put them. There wasn’t enough room in the team’s small office area in Building 8. Microsoft had been growing so fast that not only was there a shortage of big buildings on campus, but all the existing buildings were having space-crunch problems. Lill got on the phone and called a Seattle real estate company, figuring there had to be some office space available somewhere near the campus. If Marvel was going to fly, the team had to have the space in which to set up a large lab and to accommodate future growth.
Lill called Coldwell Banker, which unbeknownst to him was Microsoft’s outside realtor. Microsoft employees, even managers working on hot projects, are not supposed to look for office space on their own. A representative from Coldwell Banker phoned Microsoft’s property department, and word of what Lill had done soon filtered up to Gates. Lill got a scolding, but he also got his office space—way out on the far northeast part of the campus, in something called the East Tech building, near the company store and the executive briefing center. They took over the space left vacant when the PC repair folks moved off campus.
Lill went over to check out the building, a one-story structure with a concrete floor, a plus because of all the heavy computer equipment the Marvel lab would need. (The floor in the Windows NT lab had collapsed under the weight of the equipment, so Lill wanted the Marvel lab on the ground floor.)
“It was perfect,” said Lill. “It was off by itself. Nobody was going to bug us. There was great space for labs. We managed to get that whole building. I called it the Microsoft Enterprise Zone because it was this little sort of crappy place, but great because we had the room we needed; and frankly, I liked being away from the rest of the campus. I wanted to avoid a big political situation, where everybody wanted a finger in our pie and needed to know our plans. I really wanted to be way out there on the side so we could get this thing done and get it launched.”
Siegelman and Lill began hiring like crazy, and started setting up a lab with six server computers. In keeping with the comic-book motif, it became known as the Manga Lab because the man who ran it had spent a lot of time in Japan, where a manga is an erotic comic book.
Among the new influx of bodies was Bill Miller, the project’s marketing director. Several others came from a Seattle company called the Daily Planet, which was developing software tools for multimedia on-line content. Although the Daily Planet had initially told Siegelman it did not want to do business with Microsoft, it eventually agreed to a friendly buyout, and Microsoft hired its technical staff to work on Marvel (where else would characters from Superman’s building work?). The staff from the Daily Planet formed the core of what would become the Blackbird group, the code name for a project to develop a set of proprietary software tools for exclusive use by companies creating content for Microsoft’s on-line service. It allowed content developers a simpler way to create a unique look and feel for their services, with better graphics and interactive capabilities.
The idea for Blackbird—the code name came from the high-tech, superfast, high-flying Cold War spy plane—had grown out of a conversation that Lill and Siegelman had before the Marvel team moved into Building 8. They wanted Marvel to contain something similar to Microsoft Publisher, with sonic casy-to-usc features that would let organizations post on-line content, enabling a school district, for instance, to post its PTA minutes.
By November, though, the Marvel design had begun to shift away from its computer-user focus and toward consumers. As technical manager of the project, Lill saw some risks in this change in strategy. Although a consumer-oriented service certainly would have a wider audience, it also meant that a significantly larger database center would have to be designed into Marvel’s infrastructure. And there were continuing problems fining up publishers to post content. In those days, everyone who was anyone in the publishing business, from the Wall Street Journal to Ziff-Davis, was passing through Microsoft, usually to talk with the Exchange folks or with Craig Mundie’s group working on interactive television. They also stopped off to visit with Siegelman.
“It was pretty clear,” said one member of the Marvel team, “that we were not going to be able to provide them with really sophisticated and ready tools they could use to publish sophisticated stuff on-line anytime soon.”
Meanwhile, the marketing team was trying to come up with a name for the on-line service. A number of trade names already had been rejected. Finally, an expensive name-search firm was hired, which came up with the name ONVO. “None of us knew what the hell it meant,” said Lill. “I still don’t know what it was supposed to mean.”
ONVO didn’t mean anything, and that was the point, said a member of the project’s marketing team. “It was a word like Kodak or Xerox,” he said. “It didn’t mean anything but what you put into it with marketing and product. We liked the ‘on’ part, and the letters seemed great for logos.”
When the marketing staff met with Ballmer to pitch the ONVO name, he threw them out of his office. “Fuck, no! We are not going to call this ONVO!” he screamed. Gates had a similar response, as did a lot of others. In time, everyone agreed just to call it the Microsoft Network. (After the Microsoft Network was redesigned for the Internet in 1995,“on MSN" was used in marketing.)
New digs, new lab, more people, and now a name. The Microsoft Network was finally getting up to speed. Then, one night in early November, Russ Siegelman’s brain exploded— literally. He was at the opera when his head started to hurt. But it was no ordinary headache. He had suffered a brain aneurysm. The next morning, Siegelman’s administrative assistant called the management team together and through tears told them that Russ was in the hospital. A short staff meeting with Myhrvold followed. Within days, Siegelman had surgery. He would not return to work until the last week of December. Lill, second in command, was due to leave on a vacation to the Caribbean with his family when Siegelman fell ill. His wife and two young daughters went on without him.
The team motored on without Siegelman, almost on autopilot. They were facing a difficult, perhaps impossible schedule that required that Marvel be ready to ship with Windows 95 in June 1994. That meant beta testing would have to begin in the early part of the year. What became known as the death march had started. Members of the development team were working all hours of the day and night. Lill had meals catered in every night.
Allard, who would become one of the first within Microsoft to beat the drum for the Internet, said the message he got from Ballmer about TCP/IP was: “I don’t know what it is. I don’t want to know what it is. [But] my customers are screaming about it. Make the pain go away.” So in early 1993, Allard took charge of the development of Microsoft’s first Internet server, which would connect Microsoft to the Internet and distribute test copies of its TCP/IP to customers.
In addition to Allard’s project, Microsoft had a business arrangement with CompuServe, which sponsored an on-line forum where customers could get technical support for Microsoft products, for which Microsoft received about $1 million a year in fees. And Microsoft’s Exchange group, which was developing an e-mail and groupware product, was talking about an on-line deal with AT&T. Even the folks working on what would become BOB, the ill-fated project to develop a social interface using cartoon characters to help make applications easier to use, were talking with MCI and others about an e-mail server. (That group of managers included Gates’s wife- to-be, Melinda French.) And, of course, there was Jeff Lill’s random on-line effort, which would become the nucleus of Siegelman’s project once it got the okay front Gates.
“Russ was really doing two levels of things for Bill,” said Lill. “He was reporting on all these random efforts that were going on within Microsoft, and he was going outside the company and meeting with CompuServe and others to get a feel for what was out there before going back to Bill with his recommendations. Once Russ determined that Microsoft would not be able to buy off the big services like America Online, it was clear we would have to build our own.”
In early 1993, AOL had only about 350,000 subscribers, which made it less than a third the size of rivals Prodigy and CompuServe. But AOL boss Steve Case had decided to put the pedal to the metal to overtake his competitors. AOL began flooding the mail with free sign-up software specially designed for Microsoft’s Windows, and it went to work signing up content providers, including magazines and TV networks. New material was added almost weekly.
AOL was not the only on-line company Microsoft tried to buy as a way to jump-start its entry into the market. It also was interested in a buyout of Sierra On-Line, which was headquartered in nearby Bellevue. But that deal didn’t pan out either.
“If we could have bought something, we would have,” said Lill. “The AOL thing was for a significant amount of money. I don’t know how much, but it was significant. The initial idea was that we would just go out and buy a company and get into the service that way.”
But even though it had to start from scratch, Microsoft believed it would gain the advantage later by bundling its on-line service with Windows 95. The promise of millions of customers signing up for Microsoft’s service when they bought Windows 95 convinced Gates to give the green light. But it meant that Siegelman and his team would face a daunting schedule, because the service would have to be finished in time to ship with Windows 95, whose coming-out party was scheduled for June 1994. (That launch date would later slip to December, and finally to August 1995.)
Once Gates had given the okay, the next decision was where in the organization to put Siegelman’s development team. “Bill realized there was no real good place to put this,” said one Microsoft executive. “And there was no real division that wanted it, because it just didn’t fit with any of the existing divisions. So Bill just plunked it under Nathan. The Advanced Technology Group’s job was to provide a sort of nurturing for new stuff that wasn’t well defined enough to be in a real product unit with hard-core goals and things. So it was actually a fairly appropriate place to put this.”
After the project had been officially launched, it was time to come up with a code name, as for most Microsoft products under development. Windows 95, for example, was known in 1993 only as Chicago.
The story of how the Windows team chose to nickname Windows 95 Chicago began with a spur-of-the-moment decision in early 1992, when Jon Lazarus, then Microsoft’s vice president for strategic relations, was preparing to give a confidential briefing to software developers about an exciting new operating system that would be a successor to Windows NT. He decided to invoke an exotic locale—Cairo. The city theme immediately caught on, so when Microsoft began developing a successor to Windows 3.1 and needed a code name, a number of cities, foreign and domestic, were kicked around. “We wanted something between Seattle and Cairo in terms of functionality,” said Microsoft Vice President Brad Silverberg. “The less ambitious picked names closer to Seattle—like Spokane for a minor upgrade, all the way to London for something closer to Cairo.” Eventually, the name Cleveland was chosen. But that didn’t stick for long. “We knew Cleveland wouldn’t fly too well as a code name,” said Silverberg. A few hundred miles due west of Cleveland was that city of big shoulders, Chicago. That name stuck. Microsoft had good memories of Chicago, which had been the launch site of Windows 3.1. It was just coincidence, though a nice one, that some 375 miles south of Chicago was a city named Cairo. (Microsoft would continue the city theme when it needed a code name for the successor to Windows 95 and settled on Memphis because that was where Elvis lived. A minor upgrade was to be called Nashville, because it is near Memphis.)
Finding a code name for Microsoft’s on-line service wasn’t so complicated and had nothing to do with geography. Siegel- man’s team came up with Marvel by doing an on-line search of Prodigy’s thesaurus. The Marvel Comics name immediately fired up the team because of the prospect of naming servers and other components of the project after some of America’s superheroes, like Batman.
No Microsoft code name remains a secret for long, however; nor is it supposed to. Microsoft executives use project code names when giving confidential briefings to outsiders, and they eventually make their way into the public domain through media reports. Thus, Marvel became known outside of Microsoft early in its development. Consequently, the company received a letter from a lawyer representing Marvel Comics, who explained that the name was a trademark and Microsoft could not use it. But as far as Microsoft was concerned, the Marvel name was for internal use only and was not going to be the name of the on-line service when it came to market. The letter was ignored.
At the beginning, the Marvel team consisted of fewer than 10 people, including Siegelman and Lill. The team took over an office area in Building 8, where Gates himself worked. The team’s first mission was to come up with a plan for the kind of service they were going to build.
“We did some market research to get a feel for what the hell we were doing, for what the goal was,” said Lill. “We were going into this with the goal that it was going to be a computer user-oriented service, rather than a general-purpose service.
So, we were looking at our CompuServe traffic and saying ‘Hey, we’ve got a lot to offer a computer user. A lot of users are going to CompuServe to visit Microsoft support areas. We know something about that. So we, as Microsoft, have something to offer computer users.’ We were originally going to focus on being a very targeted product. And we did some very early prototypes that we showed some of these market research folks down in Los Angeles.”
Lill and Siegelman also flew out to Washington, D.C., to visit a tiny company that had developed some on-line technology that they thought might be useful in starting Microsoft’s service. The company was providing on-line support for software companies, with bulletin boards, e-mail, and a lot of the services that the Microsoft team thought they might want to include in their own, even though it was all UNIX-based.
Siegelman, who had a reputation around Microsoft as a first-rate negotiator, reached an agreement with the company for Microsoft to license its product, pending a technical review by Lill. But Lill recommended they pass on the license deal, even though he found it to be a very good product. He decided it just didn’t fit with the Marvel plan.
The initial design for Marvel looked a lot like Prodigy, which Lill soon realized was the wrong approach. “It was not very Windows-like,” he said. “We were trying to go for a very, very easy-to-use thing in the prototype. And just after that, I realized that we were reinventing the wheel. We would have had to create new user interfaces, and we were not a user interface group, and our overall goal was to ship something with Windows 95. And so we just didn’t have the time to be monkeying around and inventing things. And I had seen the BOB effort, and it was a black hole of changes and changes and changes and changes. So I basically told the team, ‘We are going to make this just be Windows.’ ”
The group was under the gun. Gates was supposed to be briefed on the project in late September. And he wanted more than a report; lill wanted a demonstration, proof that it could actually work if the company invested further time and resources. True, Gates had approved the project in May, but it would be the September meeting at which he would decide whether to go forward or scrap the effort. Thus the team had to be able to convince Gates at that meeting that they could do this. But could they?
A number of technical problems had to be resolved before the fateful meeting, the most important of which was to decide the underlying infrastructure for using an X-25 network. This was an older technology, but very similar to the TCP/IP that Allard was working on. AOL, Prodigy, and CompuServe all used the X-25 network for their on-line services.
The team by now was also thinking about content. What would the service offer customers? Because it was being designed primarily for computer users, the initial intent was to follow the model of CompuServe and make various computer- oriented publications available. The service would also, of course, need a traditional e-mail function and bulletin boards. The plan was to use Exchange for both these capabilities. “All that was pretty much the core technology of what we were looking at to deliver with the initial service,” said Lill. “We figured that since it was mostly for computer-oriented users, they’d be a little bit more forgiving.”
It was Lill who gave the crucial briefing to Gates at the end of September. The team managed to get a simple demo functioning on an X-25 network. They even had some image-compression technology working, and downloaded ap. image using a 9600 baud modem. “We were really just trying to show Bill that we had the technology figured out,” said Lill. “We had the scary modem figured out and the scary network figured out, and we actually had a server running and making this thing work.”
As he always does at such briefings, Gates asked some tough technical questions. But overall, the presentation went well. Not once did Gates scream at anyone—as he had been known to do at these briefings—or throw something, or shout his familiar phrase, “This is the stupidest thing I’ve ever heard of!”
Nevertheless, Gates was highly skeptical that the team would be able to meet the June 1994 shipping date for Windows 95. He didn’t think it could be done. He also had some nontechnical questions. “How much is all this going to cost?” he wanted to know. “Can we really make any money?”
Lill did not have ready answers for those questions, but Gates gave his approval anyway and authorized the hiring of some 30 more people for Marvel.
Which raised the question of where to put them. There wasn’t enough room in the team’s small office area in Building 8. Microsoft had been growing so fast that not only was there a shortage of big buildings on campus, but all the existing buildings were having space-crunch problems. Lill got on the phone and called a Seattle real estate company, figuring there had to be some office space available somewhere near the campus. If Marvel was going to fly, the team had to have the space in which to set up a large lab and to accommodate future growth.
Lill called Coldwell Banker, which unbeknownst to him was Microsoft’s outside realtor. Microsoft employees, even managers working on hot projects, are not supposed to look for office space on their own. A representative from Coldwell Banker phoned Microsoft’s property department, and word of what Lill had done soon filtered up to Gates. Lill got a scolding, but he also got his office space—way out on the far northeast part of the campus, in something called the East Tech building, near the company store and the executive briefing center. They took over the space left vacant when the PC repair folks moved off campus.
Lill went over to check out the building, a one-story structure with a concrete floor, a plus because of all the heavy computer equipment the Marvel lab would need. (The floor in the Windows NT lab had collapsed under the weight of the equipment, so Lill wanted the Marvel lab on the ground floor.)
“It was perfect,” said Lill. “It was off by itself. Nobody was going to bug us. There was great space for labs. We managed to get that whole building. I called it the Microsoft Enterprise Zone because it was this little sort of crappy place, but great because we had the room we needed; and frankly, I liked being away from the rest of the campus. I wanted to avoid a big political situation, where everybody wanted a finger in our pie and needed to know our plans. I really wanted to be way out there on the side so we could get this thing done and get it launched.”
Siegelman and Lill began hiring like crazy, and started setting up a lab with six server computers. In keeping with the comic-book motif, it became known as the Manga Lab because the man who ran it had spent a lot of time in Japan, where a manga is an erotic comic book.
Among the new influx of bodies was Bill Miller, the project’s marketing director. Several others came from a Seattle company called the Daily Planet, which was developing software tools for multimedia on-line content. Although the Daily Planet had initially told Siegelman it did not want to do business with Microsoft, it eventually agreed to a friendly buyout, and Microsoft hired its technical staff to work on Marvel (where else would characters from Superman’s building work?). The staff from the Daily Planet formed the core of what would become the Blackbird group, the code name for a project to develop a set of proprietary software tools for exclusive use by companies creating content for Microsoft’s on-line service. It allowed content developers a simpler way to create a unique look and feel for their services, with better graphics and interactive capabilities.
The idea for Blackbird—the code name came from the high-tech, superfast, high-flying Cold War spy plane—had grown out of a conversation that Lill and Siegelman had before the Marvel team moved into Building 8. They wanted Marvel to contain something similar to Microsoft Publisher, with sonic casy-to-usc features that would let organizations post on-line content, enabling a school district, for instance, to post its PTA minutes.
By November, though, the Marvel design had begun to shift away from its computer-user focus and toward consumers. As technical manager of the project, Lill saw some risks in this change in strategy. Although a consumer-oriented service certainly would have a wider audience, it also meant that a significantly larger database center would have to be designed into Marvel’s infrastructure. And there were continuing problems fining up publishers to post content. In those days, everyone who was anyone in the publishing business, from the Wall Street Journal to Ziff-Davis, was passing through Microsoft, usually to talk with the Exchange folks or with Craig Mundie’s group working on interactive television. They also stopped off to visit with Siegelman.
“It was pretty clear,” said one member of the Marvel team, “that we were not going to be able to provide them with really sophisticated and ready tools they could use to publish sophisticated stuff on-line anytime soon.”
Meanwhile, the marketing team was trying to come up with a name for the on-line service. A number of trade names already had been rejected. Finally, an expensive name-search firm was hired, which came up with the name ONVO. “None of us knew what the hell it meant,” said Lill. “I still don’t know what it was supposed to mean.”
ONVO didn’t mean anything, and that was the point, said a member of the project’s marketing team. “It was a word like Kodak or Xerox,” he said. “It didn’t mean anything but what you put into it with marketing and product. We liked the ‘on’ part, and the letters seemed great for logos.”
When the marketing staff met with Ballmer to pitch the ONVO name, he threw them out of his office. “Fuck, no! We are not going to call this ONVO!” he screamed. Gates had a similar response, as did a lot of others. In time, everyone agreed just to call it the Microsoft Network. (After the Microsoft Network was redesigned for the Internet in 1995,“on MSN" was used in marketing.)
New digs, new lab, more people, and now a name. The Microsoft Network was finally getting up to speed. Then, one night in early November, Russ Siegelman’s brain exploded— literally. He was at the opera when his head started to hurt. But it was no ordinary headache. He had suffered a brain aneurysm. The next morning, Siegelman’s administrative assistant called the management team together and through tears told them that Russ was in the hospital. A short staff meeting with Myhrvold followed. Within days, Siegelman had surgery. He would not return to work until the last week of December. Lill, second in command, was due to leave on a vacation to the Caribbean with his family when Siegelman fell ill. His wife and two young daughters went on without him.
The team motored on without Siegelman, almost on autopilot. They were facing a difficult, perhaps impossible schedule that required that Marvel be ready to ship with Windows 95 in June 1994. That meant beta testing would have to begin in the early part of the year. What became known as the death march had started. Members of the development team were working all hours of the day and night. Lill had meals catered in every night.
