第一章：梦想的种子 / Chapter 1: The Seed of Dream

2011年夏，上海。交通大学徐汇校区的宿舍楼里，热浪裹挟着蚊鸣，六个男生挤在狭小的空间。但在这个六人宿舍的角落，三个年轻人围在一台二手笔记本前，屏幕的微光映照着他们专注的脸庞。桌上散落着泡面桶、能量饮料罐，还有一张手绘的赛博朋克风格草图。那是他们梦想的起点——用2000元启动资金，在虚拟世界建造一座永不倒塌的城堡。

In the summer of 2011, in Shanghai. Inside a dormitory building at Xuhui Campus of Shanghai Jiao Tong University, heatwaves and the buzz of mosquitoes filled the air. Six boys were crammed into the tiny space. But in one corner of this six-person dorm, three young men huddled around a second-hand laptop, the screen's glow reflecting on their focused faces. Scattered on the desk were instant noodle cups, energy drink cans, and a hand-drawn sketch in cyberpunk style. It was the starting point of their dream—to build an unbreakable castle in a virtual world with a startup capital of just 2000 yuan.

这三个人，蔡浩宇、刘伟和罗宇皓，都是上海交通大学的学生。蔡浩宇主修计算机科学，代码于他而言如同呼吸般自然；刘伟钻研软件工程，擅长将疯狂的想法拆解为可执行的步骤；罗宇皓沉浸于数字媒体艺术，能用像素和多边形勾勒出令人心颤的美学。他们因共同的爱好——动漫、游戏，尤其是《魔兽世界》——而走到一起。在无数个深夜的宿舍卧谈会上，他们谈论的从来不是考试与论文，而是艾泽拉斯的史诗、二次元世界的绮丽，以及一个越来越清晰的执念：为什么中国不能有自己的顶级游戏IP？为什么我们不能创造一个让全球玩家沉浸的幻想宇宙？

The three—Cai Haoyu, Liu Wei, and Luo Yuhao—were all students at Shanghai Jiao Tong University. Cai majored in computer science; for him, coding was as natural as breathing. Liu studied software engineering, excelling at breaking wild ideas into executable steps. Luo immersed himself in digital media art, capable of crafting heart-stopping aesthetics with pixels and polygons. They were drawn together by shared passions—anime, video games, especially *World of Warcraft*. During countless late-night dorm conversations, they never talked about exams or papers. Instead, they discussed the epics of Azeroth, the beauty of two-dimensional worlds, and a growing conviction: Why couldn’t China have its own top-tier game IP? Why couldn’t they create a fantasy universe that would captivate players globally?

2000元，是三人掏空口袋凑出的全部。蔡浩宇拿出了攒了两年的奖学金，刘伟卖掉了闲置的游戏显卡，罗宇皓则用兼职设计海报的收入补足。这笔钱在当时的上海，连一台像样的开发电脑都买不起，但他们用800元淘来一台二手笔记本，剩下的钱用于租用廉价的云服务器、购买基础软件许可，以及——最重要的——维持最基本的生活。泡面成了主食，宿舍熄灯后，他们常常摸黑对着屏幕，因为舍不得开灯多耗电。

The 2000 yuan was everything they could scrape together. Cai Haoyu contributed his two-year scholarship savings, Liu Wei sold his spare graphics card, and Luo Yuhao topped it up with earnings from designing posters. In Shanghai at that time, this sum wasn’t enough to buy a decent development PC. Instead, they spent 800 yuan on a second-hand laptop, using the rest for cheap cloud server rentals, basic software licenses, and—most importantly—bare-minimum living expenses. Instant noodles became their staple. After the dorm lights went out, they often worked in the dark, reluctant to waste electricity on lighting.

他们的第一个实验品，是一款名为《泡泡枪》（Bubble Gun）的Flash小游戏。玩法简单：玩家控制一个角色，发射彩色泡泡击碎不断生成的障碍。这个项目没有商业野心，纯粹是一次技术验证——他们想测试物理引擎的手感、粒子效果的渲染效率，以及最核心的问题：在一个极度有限的资源下，能做出多“好玩”的交互？

Their first experiment was a simple Flash game called *Bubble Gun*. The gameplay was straightforward: players controlled a character shooting colorful bubbles to break continuously generated obstacles. This project had no commercial ambition; it was purely a technical proof-of-concept. They wanted to test the physics engine’s feel, the rendering efficiency of particle effects, and, most importantly: with extremely limited resources, how “fun” could an interaction be?

“我们不是要做3A大作，”蔡浩宇在日记里写道，“我们是想知道，当所有资源都归零时，创意本身能走多远。” “We’re not aiming for AAA titles,” Cai Haoyu wrote in his diary. “We want to know how far creativity alone can go when all resources are reduced to zero.”

《泡泡枪》的开发周期不到一个月。没有美工，罗宇皓就用几何图形拼凑角色；没有音效，刘伟从免费音效库中剪辑；没有服务器，游戏直接打包成单机版。但当他们在校园里邀请同学试玩时，一个意外的反馈出现了：很多人说，“这游戏有‘魂’。”——尽管画面粗糙，但操作反馈的“打击感”出奇地扎实，泡泡破裂的瞬间伴随着细微的屏幕震动和音效，竟让人产生真实的满足感。

Development for *Bubble Gun* took less than a month. With no dedicated artist, Luo Yuhao pieced together characters from basic shapes. With no sound designer, Liu Wei clipped sounds from free libraries. With no server, the game was packaged as a standalone. But when they invited classmates to try it on campus, an unexpected feedback emerged: many said, “This game has *soul*.” Despite the rough graphics, the “impact” of the operation feedback was surprisingly solid. The moment bubbles burst, accompanied by subtle screen shakes and sound effects, created a genuine sense of satisfaction.

这束微光，点燃了更大的野心。《泡泡枪》证明了一件事：在极致约束下，创意与技术的化学反应，可以超越资源的贫瘠。但Flash平台很快显露出局限——无法触及移动端，而2011年的智能手机浪潮正席卷而来。苹果App Store上线仅三年，安卓生态初具规模，他们意识到：未来属于口袋里的屏幕。

This glimmer sparked greater ambition. *Bubble Gun* proved one thing: under extreme constraints, the chemical reaction between creativity and technology could transcend resource scarcity. But the Flash platform soon revealed its limits—it couldn’t reach mobile devices. And in 2011, the smartphone wave was sweeping in. The Apple App Store was only three years old, the Android ecosystem was taking shape. They realized: the future belonged to screens in pockets.

“我们要做手机游戏，”蔡浩宇在宿舍白板上画下一个苹果图标，“但不要做那些换皮卡牌。我们要做的是……能让人心跳加速的动作游戏。” “We’re making a mobile game,” Cai Haoyu drew an Apple icon on the dorm whiteboard. “But not those reskinned card games. We’re making… an action game that gets your heart racing.”

转向移动端，意味着技术栈的全盘重写。他们选择Unity引擎——当时刚在移动平台崭露头角，学习曲线陡峭，但跨平台潜力巨大。三个几乎没接触过C#编程的学生，开始啃官方文档、看海外教程、在论坛里提问。罗宇皓负责3D建模，受限于笔记本性能，他不得不把角色面数压到极致，用贴图技巧伪造细节；刘伟设计关卡和数值，用Excel表格模拟数百次战斗平衡；蔡浩宇则主攻核心战斗系统，从零实现一个包含帧率锁定、伤害计算、受击反馈的引擎模块。

Shifting to mobile meant a complete rewrite of their tech stack. They chose the Unity engine—just emerging on mobile platforms, with a steep learning curve but huge cross-platform potential. Three students with almost no C# experience began devouring official documentation, watching overseas tutorials, and asking questions on forums. Luo Yuhao handled 3D modeling; limited by the laptop’s performance, he had to reduce polygon counts to the extreme, using texture tricks to fake details. Liu Wei designed levels and数值, simulating hundreds of combat balances using Excel spreadsheets. Cai Haoyu focused on the core combat system, implementing a module from scratch that included frame rate locking, damage calculation, and hit feedback.

资金在加速燃烧。2000元很快见底，他们不得不做出艰难选择：蔡浩宇和刘伟暂停所有兼职，全身心投入；罗宇皓接外包设计补贴开销，常常通宵赶稿后白天继续开发。宿舍成了战场，书桌被拆掉，换成两张并排的桌子，一台电脑负责编码，一台负责渲染。墙上贴满了手绘的角色设定、关卡布局图，以及一句用红笔写下的标语：“要么不做，要么做到极致。”

Funds were burning fast. The 2000 yuan ran out quickly, forcing tough choices: Cai Haoyu and Liu Wei quit all part-time jobs to focus full-time; Luo Yuhao took on outsourced design work to subsidize, often working overtime on external projects by night and developing by day. The dorm became a battlefield. Desks were removed and replaced with two side-by-side tables: one for coding, one for rendering. Walls were covered with hand-drawn character designs, level layouts, and a slogan written in red marker: “Either don’t do it, or do it to the extreme.”

2012年初，游戏雏形初现。他们给它取名《崩坏学园》（Honkai Academy）。“崩坏”一词源自日语“崩壊”（Hōkai），意为崩塌、毁灭，但他们想赋予其新的含义——在破碎的世界中，少女们挥舞武器，于废墟上重建秩序。这隐喻了他们的处境：资源匮乏如崩坏世界，但创意与热情是重建的砖石。美术风格上，罗宇皓融合了日式动漫的鲜明线条与美式漫画的硬朗阴影，创造出一种独特的“崩坏视觉”：角色身着 stylized 校服，武器既有科技感又有奇幻元素，而背景则是被“崩坏能”侵蚀的都市废墟。

By early 2012, a game prototype emerged. They named it *Honkai Academy* (“崩坏学园”). “Honkai” comes from the Japanese “崩壊” (Hōkai), meaning collapse or destruction, but they wanted to give it new meaning—in a shattered world, girls wield weapons to rebuild order from ruins. It mirrored their own situation: resource scarcity like a collapsed world, but creativity and passion were the bricks of reconstruction. Visually, Luo Yuhao blended the bold lines of Japanese anime with the tough shadows of American comics, creating a unique “Honkai aesthetic”: characters in stylized school uniforms, weapons blending tech and fantasy, set against urban ruins eroded by “Honkai energy.”

技术实现上，他们面临手机性能的严酷考验。2012年的中低端安卓机，内存仅512MB，CPU主频低于1GHz。为了在如此低的配置下维持60帧，蔡浩宇开发了一套动态分辨率系统：当场景特效增多时，自动降低渲染分辨率；当角色动作复杂时，简化骨骼计算。他还实现了“对象池”技术，避免频繁创建销毁游戏对象导致卡顿。这些后来被米哈游称为“性能优化三板斧”的雏形，此刻只是三个学生在无数个凌晨，对着帧率曲线一次次的试错与调试。

Technologically, they faced harsh constraints of mobile performance. In 2012,