@wikeygo: 藏在 WinRAR 里,俄罗斯程序员封神三十年的天才设计! WinRAR 背后的俄罗斯程序员尤金・罗沙尔,出身苏联 “坦克城” 车里雅宾斯克,早年靠着仅有 96KB 内存的仿制苹果电脑编程,从小养成极致节约字节、追求高效代码的习惯。 19…
摘要
本文讲述了WinRAR背后的俄罗斯程序员尤金·罗沙尔的故事,包括他在苏联时期的编程背景、1993年自研压缩算法、1995年推出WinRAR,以及40天试用设计的精妙商业策略:不限制功能只弹窗提醒,让个人免费使用形成生态,迫使企业批量采购授权,同时开放解压代码但封闭压缩算法以维持市场壁垒。
查看缓存全文
缓存时间: 2026/07/03 00:28
Hidden in WinRAR: The Genius Design That Made a Russian Programmer a Legend for 30 Years!
The Russian programmer behind WinRAR, Eugene Roshal, comes from the Soviet “tank city” Chelyabinsk. In his early years, he programmed on an Apple II clone with only 96KB of memory, developing an extreme habit of saving every byte and pursuing efficient code.
In 1993, at age 21, he created his own compression algorithm, surpassing ZIP in compression ratio. In 1995, he launched WinRAR and split the work with his brother: Eugene wrote the code, while his brother handled all operations, copyrights, and commercialization.
His most legendary move was the 40-day trial design, thirty years ahead of its time. This is no bug—it’s top-tier business strategy:
- If the software locked up upon expiration, everyone would switch to cracked versions, and the RAR format would completely lose the market.
- Instead, it only shows a reminder popup, doesn’t restrict any functionality, and lets individuals use it for free forever. This allowed RAR to spread across the entire internet and become a universal standard.
When entire companies rely on RAR for daily office work, copyright compliance pressure forces businesses to purchase licenses in bulk. The strategy: pave the way with individual ecosystem, profit from enterprise payments.
He also set the rules: open-source the decompression code, keep the compression algorithm closed. 7-Zip can extract but cannot create RAR files—a legal barrier to secure the market.
Eugene is extremely low-profile: no social media accounts, rarely appears in public. For 30 years, he has focused solely on polishing this tool, which is less than 4MB in size, without seeking funding or going public.
相似文章
@vintcessun: 午间看到一个解决下载目录乱象的工具,有点离谱——用Rust+Tauri写了个系统托盘文件整理器,内存占用才5MB,还刚补上Linux支持。核心是文件监控加规则引擎,按扩展名或正则自动分类,SQLite记操作历史可一键撤销。隐私方面零上报,…
介绍了一个开源的桌面文件整理工具Mouzi,基于Rust和Tauri构建,内存占用仅5MB,支持文件监控和规则引擎自动分类,强调隐私零上报。
@XAMTO_AI: 有个老哥直接把“科学上网”整成了一个开源项目,还专门针对俄罗斯那种变态级别的封锁环境来搞。 俄罗斯那边的网络管控有多离谱你们知道吗?不是简单封几个网站,而是连 VPN 协议都要一个个识别拦截。结果这位直接把能用的配置全公开扔 GitHub…
一个开源项目专门针对俄罗斯高强度网络封锁环境,提供自动检测更新、多国出口节点的VLESS VPN配置,通过GitHub免费分发。
@ArchiveExplorer: 这简直就是个纯金矿——几乎没人用过的最佳 REA 技巧。核心思路:REA 拆解应用,Jev 只保留……
介绍 REA(Reverse Engineer Anything)这一 MCP 工具及其「干净室重建」工作流:REA 逆向分析应用生成带证据的 SPEC.md,Jev 审核规则,Opus 5.5 在未见过原代码的情况下重建功能,Haiku 5.5 用故意破坏的版本跑测试,实现模型不复制原代码的合规重建。
@vintcessun: 发现新思路:编码 agent 原来可以轻到这个程度。16MB 内存、0% 空闲 CPU、26MB 二进制——这数据在 agent 圈里有点离谱。 它用 Rust 硬怼,17k LoC 就把所有标准工具、权限系统、会话管理、MCP 都塞进去…
Zerostack 是一个用 Rust 编写的极简编码 agent,仅消耗 16MB 内存和 26MB 二进制,通过 feature gate 实现轻量化,支持多种提供商和工具。
@LanLance24: Pi 发布了一篇博客详解他们的压缩机制设计,把上下文怎么膨胀、压缩怎么触发、缓存付出什么代价等进行了深入分析。
本文详细介绍了Pi编码代理的上下文压缩机制,分析了上下文膨胀原因、压缩触发条件以及缓存失效的代价,旨在优化长对话中的AI性能。