系统与基础设施ploum.net
我们能造一台设计寿命至少五十年的电脑吗？
Could we build a computer designed to last at least fifty years? (2021)
2021 — 2024，它一次次回到首页
黑点是上首页（≥ 50 分）的那次投稿，空心点是没冲上去的投稿，点越大分数越高。
导读
这篇文章探讨能否造出一台设计寿命至少五十年的电脑，作者从打字机的可修复性出发，主张简化软硬件栈、使用可自行维修的部件，让机器长期可用。文中也讨论了现代设备为何三四年就要换，以及"长寿计算机"在现实中的取舍。
为什么经久不衰：它把"电子垃圾、可维修性、技术复杂度"这些今天更紧迫的话题提前摆上台面，并且引发了对"机器是坏了还是被淘汰了"的持久争论。
HN 评论精选
真正的瓶颈不是硬件寿命，而是外部生态（加密、接口、软件支持）的淘汰，数据无法导出就等于报废。
我看到的最大问题是：就算这台电脑本身还能开机，它周围的世界也已经变得面目全非，实际上等于没用。比 50 年年轻得多的机器就能说明这点：它们支持不了现代网站要求的加密，接口也会被淘汰——你最近用过 24 针并口吗？调制解调器呢？即便是长期支持老机器的软件，最终也会发现无法继续支持更老的硬件，于是你只能停留在最后一个还能用的版本上，而厂商多半早已倒闭。机器不是因为停止工作而老去，而是因为对周围环境变得过时而废弃。文章用"你只做发邮件、文字处理这类简单用途"轻轻带过了这个问题，可如果数据无法从机器里导出，你就完蛋了。
smackeyacky查看英文原文
The major problem I see with this is that while the computer itself might still boot, the world around it will have changed so radically it will effectively be useless. You can see this with machines much younger than 50 years, they can't support the encryption required by most modern websites, interfaces age out of regular use: used a 24 pin parallel port lately? How about a modem? Even software with long support for ancient machines eventually find it impossible to support older hardware, so you would be stuck with whatever the last release of software was available and likely any vendor would be long dead. Machines don't age out because they stop working, they simply become obsolete to the surrounding environment. The article glosses over this problem by saying you could stick to simple use cases like email or word processing, but if you can't transfer the data off the machine, you are boned.
反驳原文对打字机可修性的浪漫化，并指出"被迫换机"更多是软件淘汰而非硬件损坏，且缺少做这件事的理由。
原文说：对大多数打字机而言，修复只是花时间的问题，不涉及复杂技能或工具，连最难的操作用简单的试错也能独自学会。什么？假设某个内部零件因缺润滑或卡入异物而受机械应力失效，你靠"简单试错"怎么判断是哪个零件坏了、为什么坏、原零件长什么样，最后不用专业知识和精密工具就把它修好？我猜用消费级工具和尺寸都不对的现成材料，我得花好几年才能可靠做到。原文还说我们每三四年就得换笔记本、每两年换手机、其他设备也要定期更换——这把两种"得换"混为一谈了：一种是老硬件跑软件不再顺畅（甚至跑不了）而难以完成工作，另一种是真的坏了。我并不比 1950 年代的人"更必须"每几年换最新款打字机那样"必须"换笔记本。另外我快速扫了一遍，文章没提为什么要做这件事——好玩当然可以，但真要做好，需要商业理由和大量资金。
l0b0查看英文原文
> For most typewriters, restoration is only a matter of taking the time to do it. There’s no complex skills or tools involved. Even the most difficult operations could be learned alone, by simple trial and error. What. Let's say an internal piece failed because of mechanical stress caused by lack of lubrication or a jammed foreign object. How do you figure out "by simple trial and error" which piece failed, why it failed, what the original piece was supposed to look like, and finally repair it, without expert knowledge and precision tools? At a wild guess I'd expect it'd take me a few years to do this reliably, using consumer-grade tools, from stock materials which are not already the exact right size in any dimension. > By contrast, we have to change our laptops every three or four years. Our phones every couple of years. And all other pieces of equipment (charger,router, modem,printers,…) need to be changed regularly. This is conflating "have to" in the sense that it's hard to get useful work done because software no longer runs optimally (or at all) on older hardware, and "have to" in the sense that it's literally broken. I don't "have to" change my laptop any more than someone in the 1950s "had to" change their typewriter for the latest model every few years. As a side note, a quick search through the article revealed no mention of why you'd want to do this. For fun, sure, but something like this is going to need a business case and lots of cash to be done properly.
用汽车、阿富汗装备和 Tainter 的复杂度理论，论证过度复杂的技术体系脆弱且不可持续。
技术网络如此复杂，确实让人隐约感到危险。汽车曾经简单到你可以拆开再装回去、还有不小把握装对，如今零件和电控系统复杂到根本做不到。留在阿富汗的那些装备就是例子：阿富汗政府根本用不了，因为它紧紧绑定在美国的整个基础设施和保障网络上，没有卫星和专业维修队就是废铁；而塔利班手里苏联时代的老装备反而能修、能用。Joseph Tainter 把这一现象与文明崩溃联系起来：社会每次遇到危机，往往靠让基础设施和社会系统更复杂来解决，但更复杂意味着需要更多能源、更多教育、更昂贵的维护；最终到达一个临界点，继续增加复杂度反而让事情更糟，高能耗的平衡点变得不稳定，系统只能下移到复杂度更低、能耗更少的平衡态——这往往表现为崩溃。
fallingfrog查看英文原文
It does feel dangerous, somehow, doesn't it, to have the web of technology be so complex? Even something like an automobile- once cars were simple enough that you could take them apart, put them back together, and have a decent chance of doing it right. But now the parts and the computer controlled systems are so complex that you really can't. All the equipment that got left in Afghanistan is a case in point: the Afghan government could never use it, because it was too tightly tied into the whole US infrastructure and support network. Without the satellites and the special maintenance crews it was all useless. Whereas the Taliban had antique gear left over from the Soviet Union, and it was repairable, and it worked. Joseph Tainter actually links this exact phenomenon to the collapse of civilizations. The theory is that every time a society reaches a crisis, they tend to solve it by making infrastructure and social systems more complex. But more complex means it needs more energy, more education, more expensive maintenance to run. Eventually things reach a tipping point where the social systems are so complicated that adding to the complexity actually makes things worse instead of better. At that point the high energy equilibrium point is unstable, and the system must shift down to a lower complexity equilibrium that requires less energy - oftentimes, much lower. This manifests as collapse. It would be fascinating if this desire that you see manifested often in popular culture for things to be simpler, less systematized (homesteading, anger at government and banks, calls for degrowth) was some kind of subconscious understanding that things were approaching just such a tipping point. So I get the sentiment. At the same time, our devices are really only useful because of the networks of servers and the billion dollar chip fabs that make them. You can't have something like the modern internet without all that stuff. So I don't see this really working, at least in the short term. That said, you can do a lot of interesting stuff with a low powered machine, and I'd love to play around with one. It would probably run a unix shell just fine. Not sure about emacs though..
50 年寿命并不稀奇，已有的老机器（如 C64）证明了这一点，文章把难度夸大了。
Commodore 64 是 40 年前推出的。我不确定 50 年对一个计算机来说有那么长、那么难？这篇文章写得好像 50 年是无法破译的象形文字一样。我至今还能像昨天才背过一样记得 LOAD "*",8,1，而原版 C64 今天仍然能跑。
JohnJamesRambo查看英文原文
The Commodore 64 was introduced 40 years ago. I’m not sure 50 years is that long or hard for a computer to do? This article is written like 50 years is undicipherable hieroglyphics territory. I can still remember LOAD “*”,8,1 like it was yesterday. And original Commodore 64s still run today.
历次讨论
同类还值得读
TTY 揭秘（2008）
The TTY demystified (2008)
经典文章深入剖析了 Unix/Linux 中 TTY 子系统的工作原理，包括线路规程、终端驱动、信号与控制字符等机制。
系统与基础设施上首页 11 次 · 2009–2025linusakesson.net编辑精选为什么 GNU grep 这么快
Why GNU grep is Fast
FreeBSD 邮件列表上一篇广为流传的帖子，解释 GNU grep 为什么快：核心在于 Boyer-Moore 字符串搜索算法配合精心优化的内部循环和 I/O，让程序在搜索时“…
系统与基础设施上首页 9 次 · 2010–2022lists.freebsd.orgXplain：为普通人讲解 X11
Xplain – Explaining X11 for the rest of us
《Xplain》是作者（HN 用户 Jasper_）写的一系列 X11 长文，用他在浏览器中重建的 X11 绘图模型做交互演示，讲解 X11 的绘图模型与窗口系统设计。
系统与基础设施上首页 6 次 · 2013–2020magcius.github.io编辑精选旋风教程：为 Linux 创建极小 ELF 可执行文件（1999）
A Whirlwind Tutorial on Creating Really Teensy ELF Executables for Linux (1999)
breadbox 写于 1999 年的经典教程，主题是如何为 Linux 手工构造并裁剪出体积极小的 ELF 可执行文件。
系统与基础设施上首页 6 次 · 2007–2025muppetlabs.comPostgreSQL 内部原理
The internals of PostgreSQL
《The Internals of PostgreSQL》是一本免费在线书籍，系统讲解 PostgreSQL 的内部实现，包括查询处理、并发控制、WAL、缓冲区管理等核心机制。
系统与基础设施上首页 5 次 · 2017–2025interdb.jp编辑精选