Show HN：Snaketron – 多人竞技贪吃蛇。 14年后回归
您好，HN，我刚刚完成了 Snaketron 的重写，这是一款竞技性多人贪吃蛇游戏。 （是的，是的，在 Rust 中）<p>关于这个游戏历史的一个有趣的故事：<p>早在 2012 年，我就（糟糕地）实现了 Snaketron 的第一个版本。问题在于我将游戏状态保存在客户端上，虽然我确实意识到这是不安全的，但我也不知道在不将延迟增加到无法玩的水平的情况下该怎么做。我当时还在上学，从未从事过游戏工作，之前甚至没有听说过“netcode”这个词。发布我的第一个项目的兴奋让我变得更好，我把这个安全问题放在“理论问题”桶中，然后将其交付。<p>甚至没有人会注意到，如果他们注意到了，他们就不会花时间闯入我的小贪吃蛇游戏项目，对吧？<p>WRONG！
查看英文原文
Show HN: Snaketron – Competitive multiplayer Snake. Back after 14 years
Hi HN, I just completed a rewrite of Snaketron, a competitive multiplayer Snake game. (Yes, yes, in Rust) A quick funny story about this game's history: Back in 2012 I had (poorly) implemented the first version of Snaketron. The issue with it was that I was keeping game state on the client, and while I did realize that it was insecure, I also had no idea what to do about it without increasing latency to unplayable levels. I was still in school, never worked on games, and haven't even heard of the term "netcode" before. The excitement of releasing my first pet project got the better of me, and I placed this security concern in the "theoretical issues" bucket, and shipped it. No one would even notice, and if they did, they wouldn't take the time to break into my little Snake game project, right? WRONG! Hacker News absolutely lived up to its name. While I was busy with my first experience of handling a traffic surge, scaling up my database and fixing bugs, several players found the security holes and were beginning their Snake rampage. Reports were coming in of unkillable snakes so I decided to start up a game to see if I could catch it. I loaded a 1v1 and my opponent was ... space invaders. My snake was a snake. Their "snake" was a rendering of the game Space Invaders. Needless to say they were also moving around and were invincible. I really wish I had taken a video of that, but I was floored that HN not just played, but hacked, my project is such an epic way. I always wanted to rebuild it properly, so here's V2! The architecture and motivations are quite different this time around. First, the state is server authoritative and lag is compensated for by client-state prediction. The game engine is written in Rust and compiled to WASM for the client side library. To minimize the ops burden, I also imposed a strict resiliency requirement on this project: it must be auto-scalable. So a node can join the cluster and live games will rebalance to it without disrupting gameplay. I also tried to make the new Snaketron a little more fun. The core mechanic is still, like classic Snake, collecting food without crashing. But now team matches have an objective to make competitive games more structured. There's also a Bronze to Grand Master ranking sub system (if enough people play, calibrating this will be interesting) and some fun features like boosting, combos, skins, and Play Of The Game detection. On Rust: One of my original goals for doing this in Rust was so that I could learn the language. That didn't work out as well as I thought it would. Usually I learn by building stuff, but AI got so good that I stopped writing code by hand and preferred to spend my nights & weekends working on what was more intereting to me than borrow checking errors: game design and the high availability stuff. On AI: Who knows if this would have gotten done without AI. Probably not, what with my day job and other things going on in life. It's 200k LOC (most of that is tests don't worry) but still. Anyway, here are the project links. Thank you for reading, I hope you enjoy the game and any feedback is very much appreciated! [0] https://snaketron.io [1] https://github.com/lopatin/snaketron [2] https://news.ycombinator.com/item?id=4028597 (original post from 2012)
Show HN: Snaketron – Competitive multiplayer Snake. Back after 14 years
Hi HN, I just completed a rewrite of Snaketron, a competitive multiplayer Snake game. (Yes, yes, in Rust)<p>A quick funny story about this game's history:<p>Back in 2012 I had (poorly) implemented the first version of Snaketron. The issue with it was that I was keeping game state on the client, and while I did realize that it was insecure, I also had no idea what to do about it without increasing latency to unplayable levels. I was still in school, never worked on games, and haven't even heard of the term "netcode" before. The excitement of releasing my first pet project got the better of me, and I placed this security concern in the "theoretical issues" bucket, and shipped it.<p>No one would even notice, and if they did, they wouldn't take the time to break into my little Snake game project, right?<p>WRONG! Hacker News absolutely lived up to its name. While I was busy with my first experience of handling a traffic surge, scaling up my database and fixing bugs, several players found the security holes and were beginning their Snake rampage. Reports were coming in of unkillable snakes so I decided to start up a game to see if I could catch it.<p>I loaded a 1v1 and my opponent was ... space invaders. My snake was a snake. Their "snake" was a rendering of the game Space Invaders. Needless to say they were also moving around and were invincible. I really wish I had taken a video of that, but I was floored that HN not just played, but hacked, my project is such an epic way.<p>I always wanted to rebuild it properly, so here's V2! The architecture and motivations are quite different this time around.<p>First, the state is server authoritative and lag is compensated for by client-state prediction. The game engine is written in Rust and compiled to WASM for the client side library.<p>To minimize the ops burden, I also imposed a strict resiliency requirement on this project: it must be auto-scalable. So a node can join the cluster and live games will rebalance to it without disrupting gameplay.<p>I also tried to make the new Snaketron a little more fun. The core mechanic is still, like classic Snake, collecting food without crashing. But now team matches have an objective to make competitive games more structured. There's also a Bronze to Grand Master ranking sub system (if enough people play, calibrating this will be interesting) and some fun features like boosting, combos, skins, and Play Of The Game detection.<p>On Rust: One of my original goals for doing this in Rust was so that I could learn the language. That didn't work out as well as I thought it would. Usually I learn by building stuff, but AI got so good that I stopped writing code by hand and preferred to spend my nights & weekends working on what was more intereting to me than borrow checking errors: game design and the high availability stuff.<p>On AI: Who knows if this would have gotten done without AI. Probably not, what with my day job and other things going on in life. It's 200k LOC (most of that is tests don't worry) but still.<p>Anyway, here are the project links. Thank you for reading, I hope you enjoy the game and any feedback is very much appreciated!<p>[0] <a href="https://snaketron.io" rel="nofollow">https://snaketron.io</a><p>[1] <a href="https://github.com/lopatin/snaketron" rel="nofollow">https://github.com/lopatin/snaketron</a><p>[2] <a href="https://news.ycombinator.com/item?id=4028597">https://news.ycombinator.com/item?id=4028597</a> (original post from 2012)
- 来源
- Show HN · show-hn
- 证据角色
- project_and_comment_discovery
- 时间状态
- known
- 用户判决
- 未判