0x01 中文版
翻译转载自 https://bbs.pediy.com/thread-176425.htm
在二十年前这个刚刚过去的周末，微软发布了Windows NT的第一个版本——该帝国的设备和服务所依赖的软件基础。NT已经有了一个辉煌的20年，而我敢肯定，未来的20年同样值得期待。这是一个很好的时机，让我们回顾和纪念过去。
首先，如果你不熟悉的Windows NT的发展历史，我强烈建议你阅读(希望你再次阅读)帕斯卡尔·扎卡里(G. Pascal Zachary)的“Showstoppers!”。(有印刷版[/Kindle版]、Nook Press的电子版等等)。“Showstoppers!”除了对这个不朽软件产品的产生过程进行了权威记述，还包含一些以前惊人的内幕。
这个字面上被命名为“新技术”(New Technology)的NT基于OS大师戴夫·卡特勒(Dave Cutler)的VMS，因此，人们可以多加一个字母变成WNT，或者Windows NT。当时卡特勒已经在开发VMS代号为Mica的下一代产品，打算运行在名为Prism的RISC平台上。但是当Digital公司取消了这个项目后，卡特勒便跳槽到微软，从此计算机的历史被永久地改变了！
开发NT的动力来自卡特勒的愿望：应该优于他认为过时的、低效的UNIX，而且他的梦想是就象曾经的UNIX那样有朝一日能使用二十年之久。恭喜，戴夫，你做到了！
当然，NT的成功大部分还要归功于微软联合创始人比尔·盖茨(Bill Gates)，一个精明人，不仅雇用了卡特勒和一批前Digital公司的天才们，如马克·鲁科维斯基(Mark Lucovsky，帮助创建了NTFS文件系统)，及少数“外人”，像大卫·汤普森(David Thompson，引领NT网络子系统的开发)；还留给他们自己施展的空间。
卡特勒的团队像微软内部的“臭鼬工厂”那样运作(译注：skunkworks一词源于美国洛克希德·马丁公司下属的高级技术研发部门)，基本上是一个独立的组织，可以做自己想做的任何事情。它做到了。
从这里开始了，事情会变得有点模糊了，因为我是遵从自己的记忆，而不是计算机行业那些证据确凿的历史记录。作为一个技术爱好者，我总是好奇NT和其他系统，如UNIX和NextStep(存在于20世纪90年代的早期至中期，从来没有公开给大众)。我记得拿起一份彩盒的Windows NT原始版本——当时被滑稽地称为Windows NT 3.1，与那时风头正劲的16位Windows作比较——好奇它诱人的、超值的$300标价。谁能够负担得起那样的东西？甚至质疑为什么他们会需要它？
那些日子里我对NT的理解是，它跑在传统PC硬件上，但是非常、非常地慢。我从来没有机会使用该产品，直到Windows NT 3.51(代号Daytona)的发布，这个版本提供了更好的性能和可靠性。当然，仅仅几个月后，Windows 95诞生了。微软将其从早期的Windows和NT版本过渡到一个新的基于桌面的操作系统外壳，资源管理器(Explorer)——替代了原来的程序管理器(Program Manager)和其它管理界面。
NT 4(也称为外壳更新版，因为它内核基于NT 3.51，不过使用了Windows 95的外壳和一些新的管理工具)是**常使用NT的第一个版本。的确，那时我已经写了好几年的书了，并且向读者提供了我的第一个软件，所以NT 4至少有两点是完美的：我为高等教育写过一本关于NT 4的书，并使用这个系统来编写Delphi应用程序；不象在Windows 95上，决不会因为硬件而时常崩溃。事实上，它坚如磐石。
(NT 4.0发布的当晚，我去参加一个聚会，被糊里糊涂地推入池中，这个情形不难想象。然后浑身湿淋淋地到当地软件店想购买其最终零售版，那个店员满腹狐疑地跟着我在店里逛，直至看见我打算为NT花费$100多。)
到这时，就我而言，NT代表着微软的任何东西都是好的。而且我渴望看到微软将这一技术推向主流。该公司也试图通过NT的下一个版本(Windows NT 5.0)来做到这点。事实上，我对微软这一版本的蓝图是如此激动，新建了一个叫做“Windows NT 5.0 SuperSite”的网站，这样我可以定期地为它写点什么。那个站点后来改名为“Paul Thurrott’s SuperSite for Windows”，迄今为止约有15年了。
然而，短短几个月后，事情发生了变化。首先，内部营销部门赢得了品牌之争，认为Windows名称远比NT的名称来得重要，产品被重新命名为Windows 2000。然后，为了将这个一再延期的产品尽快推向市场，微软停止了原来的计划，在此版本中将Windows 9x与NT的代码库进行了合并。反而在Windows 2000发布的同年，推出了饱受病垢Windows千禧版(Windows Millennium Edition, Me)，使过时的、基于DOS的Windows家族寿命延长了好几年。
(就在他们将期刊第一次改名为“Windows 2000杂志”时，我入伙“Windows NT杂志”。)
当然，在Windows XP中微软实现了其宏伟计划。不可否认，作为Windows产品线的核心主流，悄然地带动了NT的前进。但是，在许多方面NT的纯洁性早已消失。在NT 4.0的生命周期中，微软多余地将不安全、不成熟的IE浏览器代码库集成到操作系统内核，以抗衡Netscape，并引发了长达十年的反垄断诉讼。
但是NT依然坚持着，虽然很少有人能想起它的名字。事实上，当微软在谈到Windows是有史以来最具可塑性的软件时，实际上是在说NT，而不是古老的、基于DOS的Windows。NT内核无处不在，从Xbox到Windows Azure，甚至我口袋里的手机。堪称地球上最强大的技术之一。它也是最流行的。
具有讽刺意味的是，如今NT的仇敌和过去一样：某些UNIX的变体演变成了Android和iOS，目前地球上两个最流行的移动平台，而微软也在奋力推进基于NT的Windows 8.x和Windows Phone。现在，和过去一样，很少有人相信微软能取得市场的领导地位，甚至觉得能够提供说得过去的东西就不错了。但像以前在PC、工作站、工作组、小型服务器、服务器、数据中心和云计算市场那样，微软有自己的秘密武器。那就是NT。请一定不要忘记它。
其实，最后有个小趣闻需要考虑。戴夫·卡特勒从未离开微软，在完成某些方面的创新后，你可能还不知道——他带领微软接受了AMD64平台，使其成为行业标准的x64平台，现在所有的个人电脑都基于其上。例如，帮助创建了Azure，并且最近向Xbox One交付了NT的OS和Hyper-V技术——他现在又回到他所属的地方，Windows核心团队。是否会导致又一次NT复兴，或者卡特勒——他现在已经70岁高龄了——已做好了退休的准备？我无法知晓。但我希望，卡特勒先生在NT上再次出手的可能性足够让微软的竞争对手在晚上惊醒。
译后记： 我从邮件列表看到这篇文章，读后感慨万千，试着译出来推荐给大家。
大学里学习的第一种编程语言是FORTRAN，早已还给老师了，实在是很惭愧。那时还需要打孔员的帮助将一行行的代码变成一摞摞的卡片。
后来PC进来了，开始用CP/M(一种磁盘操作系统)，也曾在IBMPC那绿幽幽的屏幕上用过Bill Gates写的BASIC。还在APPLE
II上玩过“警察抓小偷”、“吃豆豆”。
再后来是PC-XT，黑暗的DOS出现了。要知道，见到吴晓军先生的213汉字系统是多么惊喜，赞叹求伯君先生WPS那48x48的点阵字库多么漂亮。于是学8088汇编、TURBO
C，可人家都用PASCAL？…… Windows 3.0登场了，各种震惊，可怜的小伙伴们……翻译这事儿，咱不专业，错陋之处，还请指正！
0x02 English version
English version copied from https://www.itprotoday.com/windows-78/20-years-windows-nt
Twenty years ago this past weekend, Microsoft released its first version of Windows NT, the software foundation on which the firm’s devices and services empire now rests. NT has had an amazing 20 years, and while I’m sure the next 20 will be just as interesting, this is a good time to step back and remember the past.
First, if you’re not familiar with the development of Windows NT, I strongly recommend reading (hopefully re-reading) “Showstoppers!” by G. Pascal Zachary. (It’s available in paper, Kindle, and Nook versions, among others.) Still the definitive account of the creation of a monumental software product, “Showstoppers!” includes some amazing blasts from the past.
Ostensibly named for “New Technology,” NT was OS guru Dave Cutler’s follow-up to VMS, and thus one can add one letter to each of the letters and arrive at WNT, or Windows NT. Cutler had been developing this VMS sequel, originally code-named Mica, to run on a RISC platform called Prism. But when Digital cancelled the project, Cutler jumped to Microsoft and changed computer history forever.
NT was driven by Cutler’s desire to one-up UNIX, which he saw as old-fashioned and inefficient, and his dream for the OS was that it would one day be in use for 20 years, just like UNIX. Congratulations, Dave. You did it.
Of course, much of the success for NT should also be given to Microsoft co-founder Bill Gates, who was smart enough to not just hire Cutler and a bunch of ex-Digital geniuses like Mark Lucovsky (who helped create the NTFS file system) and a handful of outsiders like David Thompson (who led the development of NT's networking subsystem), but to leave them to their own devices. Cutler’s team operated as a skunkworks of sorts within Microsoft, basically a separate organization that could do its own thing. And it did.
From here on out, things are going to get a bit fuzzy, because I’m dealing with my own memory and not well-documented computer industry history. As a technology enthusiast, I was always intrigued by NT and other systems like UNIX and NextStep that simply weren’t, in the early-to-mid-1990s, available to the masses. I recall picking up a box for the original version of Windows NT—which was hilariously named Windows NT 3.1 to match the versioning with the then-current version of 16-bit Windows—in a Best Buy and wondering at its heady $300 pricetag. Who could afford such a thing? And why would they even want it?
My understanding of NT in those days was that it ran on traditional PC hardware, but very, very slowly. I didn’t get a chance to use the product until the release of Windows NT 3.51 (code-named Daytona), a version that offered better performance and reliability. Of course, just a few months later, Windows 95 happened and with it Microsoft’s transition to a new desktop-based OS shell, Explorer, which replaced the Program Manager (and other manager interfaces) from earlier Windows and NT versions.
NT 4 (also called the Shell Update Release, because it was basically NT 3.51 with the Windows 95 shell and some new management tools) was the first version of NT I used regularly. Indeed, I had been writing books for a few years by then, and I had shipped my first software to the public, so NT 4 was perfect on two counts: I wrote a book about NT 4 for higher education and used this system to write Delphi applications; unlike with Windows 95, it wouldn’t crash down to the hardware regularly. In fact, it was rock-solid.
(The night NT 4.0 was released, I went to a party and was pushed into a pool in a bit of silliness that’s probably not hard to imagine. I visited the local Software Etc. to purchase the final retail version of the software soaking wet, and was followed around the store suspiciously until the employee saw I was going to spend over $100 on NT.)
By this time, NT stood for everything that was right at Microsoft, as far as I was concerned. And I was eager to see Microsoft move this technology to the mainstream. The firm intended to do just that with the next release of NT, called Windows NT 5.0. In fact, I was so impressed by Microsoft’s plans for this release that I started a new website called Windows NT 5.0 SuperSite so that I could write about it regularly. That site, since renamed to Paul Thurrott’s SuperSite for Windows, is still around 15 years later.
Just a few months later, however, things changed. First, internal marketing won the branding war and, deciding that the Windows name was far more important than the NT name, rebranded the product as Windows 2000. Then, in order to get the continually delayed product out the door, Microsoft stepped back from its plans to merge the Windows 9x codebase with NT in this release. Instead, it shipped the reviled Windows Millennium Edition (Me) in the same year as Windows 2000, extending the life of the old DOS-based Windows family another few years.
(I came onboard with Windows NT Magazine just before they changed the name of the publication, first to Windows 2000 Magazine.)
Microsoft of course realized its grand plans with Windows XP, which did indeed carry NT forward, silently, as the sole mainstream Windows product line. But in many ways the purity of NT had long since ended. During NT 4.0’s lifetime, Microsoft had integrated the unsafe and immature Internet Explorer codebase into the OS core unnecessarily and in order to harm Netscape, triggering a decade of antitrust issues.
But NT soldiered on, though few remembered its name. Indeed, when Microsoft talks about Windows being the most malleable software ever created, it’s really talking about NT, and not the old DOS-based Windows. The NT kernel lives in everything from the Xbox to Windows Azure to the phone in my pocket. It is indeed one of the most versatile technologies on earth. It is also one of the most popular.
Ironically, NT’s foes today are the same as in the past: Some variant of UNIX drives both Android and iOS, the two most popular mobile platforms on earth, while Microsoft is pushing forward with the NT-based Windows 8.x and Windows Phone. Now, as in the past, few believe that Microsoft can oust the market leaders, or even offer something passingly good. But as it did before in the PC, workstation, workgroup, small server, server, datacenter, and cloud computing markets, Microsoft has a secret weapon. It’s called NT. And don’t you forget it.
Actually, here’s one last little tidbit to consider. Dave Cutler never left Microsoft, and after innovating in some ways you might not be aware of—he spearheaded Microsoft’s acceptance of the AMD64 platform, making it the industry standard x64 platform on which all PCs are now based, for example, helped create Azure, and most recently delivered NT’s OS and Hyper-V technologies to Xbox One—he’s now back where he belongs, in the core Windows team. Will this lead to yet another NT renaissance, or is Cutler—he’s over 70 years old now—ready for retirement? I don’t know. But I hope the possibility of Mr. Cutler going hands-on with NT again is keeping Microsoft’s competitors up at night.