Pioneering the Multiplayer Revolution
Yoo Jae-won arrived at the ID Software headquarters housed within the High Performance Computer building and was immediately escorted to the twelfth-floor conference hall. The room was vast enough to accommodate over a hundred people at once, its seating arranged in a semicircle facing the central stage. A large screen dominated the stage, and the chairs and desks were of the highest quality. It was not a facility used often. ID Group operated in teams of no more than eight, and ID Software was no exception. Under normal circumstances, a quick meeting within the team was all that was needed. The grand conference hall was reserved for announcements, concepts, or critical information that had to be shared with the entire organization.
Today, the hall was filled with people for the first time in a long while. Among them were the founding members who had started ID Software in a tiny garage, though most faces were new. It seemed every employee had gathered because Yoo Jae-won was coming. Of course, ID Software was a much larger company now, with just over three hundred employees, so not everyone was present. Still, every planner, programmer, and designer involved in software development had assembled. The reason was obvious: the visit of their owner.
“The owner of ID Software and the great and sole leader of ID Group, Chairman Yoo Jae-won, is entering. Please welcome him with applause!”
What was this atmosphere? As Yoo Jae-won stepped into the hall, a title straight out of North Korea rang out, followed by thunderous applause and cheers. His entrance was even being projected in giant scale on the front screen. John Carmack let out a startled gasp. He, too, seemed caught off guard by the surprise event.
“We’ve been studying Eastern philosophy and culture at the company level to prepare for next-generation game development. I guess someone watched something strange yesterday!”
John Carmack’s remark suggested the presentation materials had been pulled from old Korean newsreels or 1970s propaganda footage. Yoo Jae-won nodded at the comment and raised a hand in acknowledgment. The applause and cheers grew even louder.
“Chairman, please say a few words.”
John Romero, a founding member of ID Software and now its vice president, who had been acting as emcee and whipping the crowd into a frenzy, respectfully handed over the microphone. Yoo Jae-won had originally come to Dallas to resolve internal disagreements at ID Software and finalize the libraries for the next version of Glide X. An impromptu speech to the entire staff had not been on the schedule, but it did not seem like a bad idea. With the launch of Nextcomcast, the internet big bang had begun, and though progress still felt slow to him, the internet was spreading far faster than it had in the original timeline. Once South Korea’s Kookmin PC initiative started, the internet would instantly become the hottest topic in the computer industry. Yoo Jae-won had been pondering killer titles for the internet for some time; brainstorming with ID Software might yield excellent results.
Taking the microphone from John Romero, Yoo Jae-won stepped onto the stage.
“Shall we talk a little about the future of the internet and games?”
Amid the enthusiastic cheers of the ID Software employees, Yoo Jae-won began his short lecture. For thirty minutes he held the conference hall spellbound, explaining the internet and games in simple, engaging terms. To someone with decades of game experience who had personally witnessed the internet’s evolution, it was hardly difficult. The core message was that while traditional games had been a confrontation between computer and human, in the internet age they would become a confrontation between human and human.
Single-player games that pitted the player against a computer inevitably contained patterns. Once computers surpassed the singularity they might generate variables as unpredictable as any human, but current technology made that impossible. Even the best-written game scripts eventually revealed their patterns after repeated play. Peer-to-peer games connected through the internet completely transcended those patterns. The moment players were matched as allies or enemies, countless variables emerged. That was why Doom 2’s multiplayer still enjoyed such fervent popularity. Furthermore, an ELO rating-based ranking system that intensified competition acted like an infinite power plant constantly fueling players. Those who disliked fierce competition could simply play free matches without ELO points; in fact, about ten percent of Doom 2’s multiplayer participants only played free matches.
Yoo Jae-won stated that any game hoping to earn long-term affection from gamers would need to be even more dedicated to multiplayer than Doom 2. He also spoke about role-playing games in which hundreds or thousands of players could simultaneously connect to a single server and assume various roles. It was only natural that John Carmack and the rest of the ID Software staff were astonished. After all, even Doom 2’s sixteen-player multiplayer servers had been limited to high-performance computers, and smooth gameplay had still been rare. Network latency, netcode optimization, and server response times created a complex set of problems. The only real solution had been the same one demonstrated at the Coppermine event: direct serial connections to large-scale servers. Yet here was someone casually proposing hundreds or thousands of simultaneous users on a single server.
John Carmack, listening from a corner of the hall, clicked his tongue in amazement while feeling a surge of competitive spirit. When one of the developers remarked that even with HPC-class machines it seemed impossible to implement, Yoo Jae-won had lightly replied that he himself could make it work. For John Carmack, who prided himself on never falling behind in programming, the comment was a powerful stimulus. His specialty was game engines, and his grand ambition was photorealistic 3D graphics, but even the upcoming HPC would struggle with that. It was a challenge that would last until the end of his development career. On the other hand, writing server programs and netcode for hundreds of concurrent users felt like something he could tackle right away. As the game engine continued to advance, it was time to add multiplayer functionality anyway; once built, it could be reused for years.
After the passionate lecture, Yoo Jae-won moved to John Carmack’s office. For a company that had earned hundreds of millions of dollars from games, the president’s office was remarkably ordinary. The space was just over twenty-six square meters, furnished with a standard desk and several development computers. Both walls were lined with library documents and magazines showcasing the latest computer technology trends. The only notable feature was the floor-to-ceiling glass window behind the desk that offered a clear view of downtown Dallas—the sole privilege of the president visible in such a cramped office.
“I once tried decorating it like one of those penthouse offices you see in magazines, but it was too distracting and I couldn’t get any work done.”
Yoo Jae-won nodded in understanding at John Carmack’s explanation. The office might look impressive, but in practice it seemed to hinder efficiency. Still, the romantic image of a grand building remained firmly lodged in Yoo Jae-won’s mind.
John Carmack picked up a remote from the desk and pressed a button. Thick blackout curtains descended, covering the glass wall. The pearly white material suggested the curtains also served as a projection screen. The lights dimmed automatically and the projector turned on. With a single remote, the entire projection system had been perfectly configured.
“Wow, everything sets up the moment the boss arrives?”
It was impressive, though the system apparently had occasional glitches. The auxiliary shell book on the desk finished booting and connected to the projector, displaying a large image.
“These are all the proposals we’ve collected so far.”
John Carmack moved the mouse to open a directory filled with dozens of IDW and IDP documents. Each file appeared to be a proposal.
“I removed the ones too embarrassing to show the boss. This is what’s left.”
Yoo Jae-won had thought the number was already large, but John explained that there had originally been far more; he had reviewed them all and discarded anything below standard.
“What do you want to make, John?”
Before looking at the proposals, Yoo Jae-won asked about John Carmack’s own intentions. The massive investment in ID Software had been made solely because of John Carmack. Thanks to that, ID Software had achieved great success from Wolfenstein through Doom 2, and Yoo Jae-won had received clear returns on his investment. John Carmack’s life had changed dramatically, as had the lives of the founding members. There had been no internal splits caused by creative differences, and they had all become multimillionaires together. With such upheaval, it was possible his mindset had shifted. He might even declare he wanted to build rockets, as he had in the previous life. After all, Doom 2’s commercial success had been enormous. It had been bundled with most 3D accelerator cards and continued selling strongly in retail. Although Aegis Shield, Yoo Jae-won’s anti-piracy program, had long since been cracked, Doom 2 remained unaffected. Its core was multiplayer; connecting to the server verified the user’s key, so a legitimate copy was required.
“Hmm, I like this one.”
The document John Carmack stopped on was titled Project KW 2. He double-clicked to open it, allowing Yoo Jae-won to view the contents on the projector.
“Ah.”
A small exclamation escaped Yoo Jae-won. It was a proposal for the official sequel to Keyboard Warrior—the game that had brought him to this position. Its novel combination of keyboard practice and zombie-killing gameplay had conquered the world when the 286 was at its peak. It had been released in Korean and English, translated into languages worldwide, and even a Chinese version had appeared. Novel and film rights had been sold as well. Because it contained a surprisingly profound story for a zombie game, the novel had sold quite well. The film adaptation, however, remained silent. Although the graphics and game system had not changed, a patch had been released for Android 2.0, and people were still using it in schools and homes.
“Wouldn’t it be fun to make Keyboard Warrior with graphics at the level of a Doom 2 expansion pack? Plus, multiplayer should be possible.”
Yoo Jae-won agreed with John Carmack. Several years had passed since Keyboard Warrior’s release. Its graphics had been impressive then, but now they looked hopelessly outdated.
“Sounds good. But this is still too small-scale. Show me the AAA proposals as well.”
This project could be handled by a single eight-person team. Yoo Jae-won had come today to select an AAA game that would require the entire ID Software development staff. In the 3D game era, development risks had grown significantly. Modeling was standard, and some titles even included cinematic cutscenes. Companies like Origin had pioneered the approach with space-opera games such as Wing Commander, casting famous actors from Star Wars and compositing them into CG cinematic sequences. Electronic Arts, having missed Doom 2, had carved a path through sports games. By securing official licenses for major sports like FIFA, NFL, and NBA, they placed real player names inside the games. Adding multiplayer features that had never existed before raised the overall quality. Of course, licensing fees on top of production costs drove total development expenses even higher. When such heavily invested games failed, the losses were devastating. Yet many studios still attempted them. The North American and global game markets were expanding daily, and the industry’s scale had grown accordingly. What had once been a matter of investing one hundred won to earn two hundred had become an era where investing ten thousand won could return thirty thousand, fifty thousand, or even a hundred thousand.
John Carmack grinned and opened three documents at once. He began explaining the first one that appeared.
“Return to Castle Wolfenstein. It’s the sequel to Wolfenstein. Single-player has been strengthened, set in a medieval backdrop with Nazis and zombies as the main enemies. We haven’t neglected the current trend of multiplayer, though. Players can enjoy deathmatch and team deathmatch using the various maps from Return to Castle Wolfenstein, and user maps will also be supported.”
“Ah, this one came up before, didn’t it? It’s been fleshed out quite a bit since then.”
A similar meeting had taken place previously. Opinions had been sharply divided between Wolfenstein and Doom 2 as the next project, and when Yoo Jae-won sided with Doom 2, the Wolfenstein proposal had been shelved.
“Yes. Romero took the loss hard and has been preparing ever since.”
John displayed the concept art included in the proposal. He had apparently worked on it in his spare time; the quality was on an entirely different level from what Yoo Jae-won had seen years earlier. The volume was also substantial. Completing single-player on normal difficulty would take at least eight hours, and twelve multiplayer maps were planned. At that scale, releasing a four-hour, six-map expansion pack separately would have been feasible, but in 1993 the culture of fragmenting a complete game for sale did not yet exist. Yoo Jae-won felt the same way; he had no intention of muddying the waters of a still-growing game industry.
“What about the others?”
“This one doesn’t have nearly as much content as Romero’s.”
As John explained, the displayed document changed. Unlike the richly detailed Return to Castle Wolfenstein, this proposal was extremely thin. The attached images were merely scanned pencil sketches. Yet the single idea it contained was undeniably powerful.
“Quake?”
Even as he read the title, Yoo Jae-won had to work to hide his inner shock. He had not expected it to appear so soon.
“It’s Team Willits’s proposal. He said he drew great inspiration from Doom 2’s multiplayer.”
Team Willits was also a founding member of ID Software and served as lead game designer on the development team.
“It’s a rather unusual game. It does have single-player, but only as practice against bots on multiplayer maps. The core is multiplayer. That’s why there’s no character progression element either.”
Yoo Jae-won nodded at John Carmack’s explanation, yet felt even more astonished inside. In the original timeline, Quake had included single-player missions. Now a version that had eliminated single-player missions from the very beginning was being proposed first. A multiplayer-only Quake had not appeared until 1999; releasing it in 1993 would put it six years ahead of schedule. Yoo Jae-won flinched for a moment, then quickly regained his composure. It was still only a proposal, not an actual title, so there was no need to panic. Besides, given Team Willits’s tendencies, a multiplayer-focused Quake emerging early was entirely plausible. The difference in game philosophy between Team Willits and Romero was clearly visible in the two proposals now displayed. Romero’s Return to Castle Wolfenstein was packed with RPG elements—characters could grow and weapons could be upgraded. Team Willits’s Quake, by contrast, was nothing but combat. Victory depended solely on player skill; there were no performance differences between weapons or armor provided on the maps.
“Each match lasts five minutes at most, ten at the longest, and it aims for extremely fast-paced combat. In exchange, it will pursue the absolute best in multiplayer and graphics.”
The external polish of the proposal resembled a first report submitted by a new college student. Yet the idea contained within was worthy of both ID Software and ID Group.
“This one’s good too!”
Yoo Jae-won gave his honest impression. John Carmack looked slightly troubled. He believed both ideas were excellent. Moreover, as a game-engine programmer, John had little interest in game design itself. That was why, when recruiting founding members, he had sought out level designers and planners—people like Romero and Team Willits. Thanks to them, he now faced a difficult choice: whether to back Romero, who had been passed over once before, or Team Willits, who had brought a fresh concept.
Yoo Jae-won’s conclusion was simple.
“Let’s do both.”
“Both of them?”
“Yes. We’ll just need to adjust the staffing. If we rate ID Software’s development capacity at ten, we should allocate eight to the larger-volume Wolfenstein and about two to Quake. Instead, we’ll actively outsource resource production such as background music, sound effects, and modeling.”
“Even so, developing two game engines at the same time…”
“One should be enough. Even though the two games look completely different in character, they share many common elements. The multiplayer portions, for example, would have a great deal of overlap.”
John Carmack had not failed to consider developing both simultaneously. Creating two fundamentally different game engines at once would have been overwhelming even for him. Yet if they limited the scope to multiplayer, there appeared to be sufficient shared components. The internal politics that had previously clouded the next-project decision suddenly felt trivial. He was surprised such a straightforward solution existed. Columbus’s egg had seemed obvious only after it was shown; at first it had looked like an impossible puzzle.
“By the way, what about the next version of Glide X?”
Having solved the immediate issue, Yoo Jae-won moved to the next topic.
“Ah, that’s about the new entrants.”
John Carmack’s expression had grown noticeably more relaxed as he began his explanation. 3D accelerator cards had become essential PC components and, consequently, a highly profitable business. Where money flowed, new companies naturally followed, and the 3D accelerator market was no different.
“Newcomers like 3DFX and NVIDIA want to join the Glide X standard, but existing companies are blocking them.”
3DFX? NVIDIA? Suddenly, names from the past came flooding back.