add uboot and vanet

Zim/Interview/AWK之父访谈录.txt

@@ -0,0 +1,112 @@
+Content-Type: text/x-zim-wiki
+Wiki-Format: zim 0.4
+Creation-Date: 2012-08-28T18:09:54+08:00
+
+====== AWK之父访谈录 ======
+Created Tuesday 28 August 2012
+{{./1}}
+      Alfred Aho
+        
+计算机科学家、编译器专家 Alfred V. Aho 一直紧盯计算机科学研究前沿。Aho 教授一直致力于**编程语言开发**，曾任贝尔实验室计算机科学研究中心副主任，现为哥伦比亚大学 Lawrence Gussman 计算机科学系教授。
+
+　　Aho 教授不仅是“龙书”系列的作者之一，而且在上世纪70年代同 Brain Kernighan 和 Peter Weinberger 一起开发了模式匹配语言 AWK。
+
+　　近日，《计算机世界》有幸邀请到 Aho 教授谈一谈 AWK 的开发。
+
+**问：您为什么要开发 AWK 语言？**
+
+　　和大部分语言一样，AWK 是**应实际需要而生**的。上世纪80年代初，我还是贝尔实验室的研究员。那时我需要追踪经费预算。同时因为我在临近的一所大学兼职，追踪学生的成绩也成了一件棘手的事情。
+
+　　我想要一种小巧的语言，能让我只写一两行代码就完成这些工作。碰巧 Brian Kernighan 也有这种需求，于是我俩一起开发了**一个专门针对简易数据处理的模式匹配语言**。
+
+　　GREP 对我们的影响很大，它是 UNIX 下一个很流行的字串匹配工具。GREP 也是我们研究中心开发的。GREP 可以根据某个正则表达式查找文本文件中特定的行，然后输出它们。
+
+　　我们想进一步扩展它，使得数字也能向字串那样处理。我们还想在打印之外提供一些计算功能。
+
+　　开发 AWK 仅仅是用来满足我们，或是那些对计算机不是很懂的人，处理常规数据的需要。它采取“**匹配模式——执行动作**”的方式工作。
+
+**问：在你们开发 AWK 时有没有哪种程序或是语言已经具备了这些功能？**
+
+　　我们最初是以 GREP 为原型（开发 AWK 的）。但是 GREP 在“//模式——动作//”处理上有一些局限，所以我们想扩展它。当时，我还在研究字串匹配算法和编译器里的上下文无关文法，自觉不自觉地借鉴了一些 LEX 和 YACC（构造编译器的工具）里的东西。
+
+　　LEX 用于**词法分析**，而 YACC 用于**语法分析**。它俩都是构造编译器的重要工具，当时在贝尔实验室很流行。后来流传开来，用来开发了很多小语言。Brian Kernighan 当时用它们来开发数学排版和图形处理语言。
+
+　　LEX 用来分拣输入文本中的词法单元（词元）。词元是一串可以构成逻辑含义的字符，比如，编程语言中的关键字“then”就是一个词元。我们对字母‘t’不感兴趣，对‘h’也不感兴趣，我们感兴趣的是‘then’这个组合。编译器的第一个部分就是__词法处理单元__，它读入源程序，分析出其中的词元。
+
+　　AWK 受这种词法分析方式的影响（很大），但 __AWK 的定位是数据处理__，而且对用户的计算机背景要求很少。
+
+**问：您能简短地向《计算机世界》的读者们介绍一下 AWK 语言吗？**
+
+　　AWK 是一门处理文本文件的语言。它把文件看作一串记录(record)，缺省情况下一行即为一个记录。每一行又被拆成若干域(field)。我们可以把一行中的第一个词看作第一域，第二个词看作第二域，以此类推。一个 AWK 程序就是一连串“模式——动作”语句。AWK 一次读入一行，然后对照程序中的各个模式进行扫描。一旦匹配成功就执行相应的操作(action)。
+
+　　举个例可能更清楚一些。假设我们有一个文件，它的每一行都是一个名字后跟一个电话号码。我们假设其中一行为“Pope     15193741273”。AWK 语言把第一域记作$1，第二域记作$2。现在我们想查 Pope 的电话号码，只需一行 AWK 语句：$1 == "Pope" { print $2 }
+
+　　这条语句的意思是：若我们找到一行，其第一域为 Pope，则输出其第二域（即电话号码）。现在你已经算是 AWK 程序员了^_^
+
+　　AWK 程序一般由一组“模式——动作”语句构成。模式可以是字串或数字的真假判断；动作是一组类似C语言的语句。
+
+　　AWK 随着成为 UNIX 的标配而流行起来。
+
+**问：在 AWK 的开发过程中，您最得意的是什么？**
+
+　　AWK 是由 Brian Kernighan、Peter Weinberger 和我三个人开发的。那时，Peter Weinberger 很好奇 Brian 和我整天在忙什么。我们写好了 AWK 的语法规范，但缺少一个完整的运行环境。Weinberger 跑来对我们说“这看起来很像我写过的一个语言”，然后就在一个星期内写好了 AWK 的运行环境。我们用这个原始版本来处理我们感兴趣的数据，很称手。更重要的是，它为 AWK 提供了一个可以不断扩展的平台。
+
+　　对我来说，最重要的是__这个项目让我了解了 Kernighan 和 Weinberger 是怎么考虑语言设计的__——这真是一次很受益的合作！有了这个灵巧的**编译器构造工具**，我们就可以完全掌控我们的开发了。我们很快加入了一些新的语法来改进 AWK。整整一年，我们都在激烈地讨论哪些特性应该加进 AWK，哪些应该丢弃。
+
+　　语言设计是一项非常随性的活动，每个设计者都会把他们需要的特性加入到语言之中，这些特性或者来自他们**需要解决的问题**，或者来自他们**解决问题的方式**。在开发 AWK 的过程中，我得到了许多乐趣，而和 Kernighan、Weinberger 他们共事是我的职业生涯中最刺激的事情。我可不想和他们那样的设计者竞争。__他们的编程能力无与伦比。__
+
+　　说来也有意思，最开始我们并没有想到，除我们三个以外还会有人用它。但很快我们发现，很多人有数据处理的需求，而这正是 AWK 所擅长的。人们不愿写几百行的C程序来做数据处理，而同样的工作只需几行 AWK 代码。于是越来越多的人开始使用 AWK。
+
+　　很多年后，AWK 仍然是 UNIX 下很常用的一个命令。时至今日，即使已经出现了一大批类似的语言，但 AWK 仍然常年保持在编程语言流行排行榜上25至30名的样子。而这一切都起于我们仨为了满足自己的需要开发的一个小工具。
+
+**问：AWK 如此流行，您有何感触？**
+
+　　我很高兴有人喜欢 AWK。AWK 不仅吸引了很多人使用它，后来很多语言的设计者也借鉴了它的工作方式。
+
+　　在 AWK 开发出来10年后，Larry Wall 开发了一门叫 Perl 的语言，借鉴了 AWK 和其他一些 UNIX 工具的特性。Perl 现在是世界上最流行的语言之一。所以不光是它刚出现的时候受欢迎，它也影响了很多新生语言。
+
+**问：您刚才提到 AWK 影响了很多语言，您认为这是为什么？**
+
+　　最初让 AWK 得以流行是__因为它的简洁和它的定位__。它有一个__很简单的编程模型__。它这种“识别模式——执行动作”的编程方式很容易理解。我们还在 AWK 中添加了对管道的支持。AWK 里的动作其实就是一些简单的C程序。你既可以写下像{ print $2 } 这样简单的动作，也可以为某个模式写一个复杂得多的类C程序。甚至华尔街的一些金融机构都用 AWK 来做账，因为它处理起数据来真的很简单。
+
+　　AWK 的学习曲线很平缓，很多人都是从 AWK 开始接触编程这行的。即使是今天，仍有很多人还在使用 AWK，他们说像 Perl 这样的语言变得太复杂了。
+
+　　AWK 的另一个优势在于__它很稳定__。自上世纪80年代中期以后，我们就没有改动过它了。所以也有很多人将 AWK 移植到不同的平台（比如 Windows）上。
+
+**问：A、W、K这个顺序最初是怎么定的？**
+
+　　这其实不是由我们定的。研究中心的同事看到我们仨挤在一个屋里，就推门喊到“AWK! AWK!”。我们觉得不错，就叫它 AWK 了。我们出版那本《AWK 编程语言》时还专门在封面里加了一只海雀（译注：AUK，音同 AWK）。                                                                              
+
+**问：您开发 AWK 时收获哪些受益至今的经验？**
+
+　　我的研究方向包括算法和程序语言。很多人知道我是因为 AWK。很多人在学术论文里引用我们的算法（已经在很多工具里实现了），但很少有人因为这个知道我。在我们开发 AWK 时借鉴了很多高效的字串匹配算法。UNIX 的很多工具，包括 EGREP 和 FGREP 这两个我研究字串匹配时写的小程序，也用到了这些算法。
+
+　　__AWK 是理论与实践结合的典范。__最好的工程实践往往是建立在科学的基础上。我们在 AWK 中引入了富于表达的记法和高效的算法。实际中，它们的效率也很高。
+
+　　__和聪明人共事你会获得智慧。__Brian Kernighan 就是这样的聪明人，一个程序语言设计大师。__他的心得就是保持简洁__，简洁的语言易于理解，也易于使用。我想，这就是对所有语言设计者最好的忠告。
+
+**问：AWK 开发出来这么多年，您有没有遇到过什么惊喜？**
+
+　　有次星期一一上班，我发现有个人在我办公室。他是贝尔实验室微电子产品部门的，他说他用几千行 AWK 写了一个计算机辅助设计系统。我当时就惊呆了。我原以为没人会写超过一千行的 AWK。但他用 AWK 写了一个很强大的 CAD 开发系统。他说用 AWK 来开发很快，很灵巧。最让我惊奇的是，我们从没想到会出现这么多用 AWK 写的应用。可能这是一个好工具的标志，就像起子不仅仅只能拧螺丝一样。
+
+**问：您今天还在用 AWK 吗？**
+
+　　我现在还在用 AWK 处理数据，它很好用。举个例子，我写论文、写报告时会用它来做图片和示例的自动编号，这需要用到它的关联数组。比如说，它可以把“图片 AWK 程序”这样的注释转换为“图片 1.1”。我写了两行 AWK 来实现这个功能。我曾见过一片论文用来1000行C代码写了一个功能还不及我那两行 AWK 的程序。用过 AWK，你就会对它在表达上的简洁干练印象深刻。
+
+**问：AWK 之父的声誉对您的职业生涯有什么影响吗？**
+
+　　我说过，很多人知道我是因为 AWK，但在计算机科学这个研究圈子里人们对我在理论方面的工作更熟悉。所以最初我只是__把开发 AWK 视为一种实践__，而不是正儿八经的研究。但开发 AWK 的这段经历，对我日后教授编程语言、编译器、软件工程这些课程还是产生了很大的影响。
+
+　　我注意到，__人们知道某些科学家不是因为他们的研究成功，反而是因为他们发明的某个工具__。比如，世界上最著名的计算机科学家——Don Knuth，计算机算法领域的奠基人。他发明了一种文字排版的语言，Tex。文字排版不是他研究的主要方向，但 Tex 却得到很多非计算机专业的科学家的广泛使用。而 Knuth 之所以要发明一套数学排版系统只是因为他希望自己的论文和著作好看一些。
+
+　　很多计算机科学家在研究之外发明了很多好用的编程语言。另一个著名的例子就是 Bjarne Stroustrup 开发 C++ 最初只是为了写一个网络模拟器。
+
+**问：如果现在重新开发 AWK，您会做出哪些改变？**
+
+　　我会在__开发之初就制定严格的测试__。我们最初把 AWK 定位为**一种用过即扔的语言**，所以在最初的实现中没有做严格的质量控制。
+
+　　我之前提到过一个用 AWK 写了 CAD 系统的家伙。他最初来见我的原因就是为了报告一个 AWK 编译器里的 bug。他很生气地对我说，他浪费了三周时间来查错，却发现原来是编译器的问题！我和 Brian Kernighan 商定，__一定要引入质量控制__。所以我们为所有的特性制定了一套严格的回归测试。打那以后，__每次我们要给 AWK 添加什么特性总会先写测试。__
+
+　　多年来，我一直在哥伦比亚大学教程序语言和编译。这门课有一个工程实践，学生们四个或五个一组用一整学期来设计他们自己的小型语言，还要做出一个编译器。
+
+　　上这门课的学生之前都没有深入了解过编译器，但这些年来从没有一个组没能完成任务。这都要归功于我和 Kernighan、Weinberger 他们一起在开发 AWK 时积攒的经验。除了学习程序语言和编译器设计的__原则__，学生们还能（通过工程实践）了解到实际中的软件开发。__每个学生都应该从一开始就养成严格测试的习惯。__学生们还能学到工程管理、团队合作、口头交流和文档撰写方面的技巧。所以从这个角度讲，AWK 真的对我教程序语言、编译和软件开发都有很深的影响。

Zim/Interview/The_History_of_the_Floppy_Disk.txt

@@ -0,0 +1,68 @@
+Content-Type: text/x-zim-wiki
+Wiki-Format: zim 0.4
+Creation-Date: 2012-08-29T19:03:08+08:00
+
+====== The History of the Floppy Disk ======
+Created Wednesday 29 August 2012
+http://h30565.www3.hp.com/t5/Feature-Articles/The-History-of-the-Floppy-Disk/ba-p/6434
+
+**The History of the Floppy Disk**
+by Steven Vaughan-Nichols (sjvn01) on 27-08-2012 08:00 AM - last edited on 27-08-2012 08:14 AM by  	 estherschindler
+
+In the fall of 1977, I experimented with a newfangled PC, a Radio Shack TRS-80. For data storage it used—I kid you not—**a cassette tape player**. Tape had a long history with computing; I had used the IBM 2420 **9-track tape system** on IBM 360/370 mainframes to load software and to back-up data. Magnetic tape was common for storage in pre-personal computing days, but it had two main annoyances: it held **tiny amounts** of data, and it was **slower** than a slug on a cold spring morning. There had to be something better, for those of us excited about technology. And there was: the floppy disk.
+{{./1.jpg}}
+//Welcome to the floppy disk family: 8”, 5.25” and 3.5”//
+
+In the mid-70s I had heard about floppy drives, but they were expensive, exotic equipment. I didn't know that IBM had decided as early as 1967 that tape-drives, while fine for back-ups, simply weren't good enough to load software on mainframes. So it was that Alan Shugart assigned David L. Noble to lead the development of “a reliable and inexpensive system for loading **microcode** into the IBM System/370 mainframes using a process called **Initial Control Program Load (ICPL).**” From this project came the first 8-inch floppy disk.
+
+Oh yes, before the 5.25-inch drives you remember were the 8-inch floppy. By 1978 I was using them on mainframes; later I would use them on Online Computer Library Center (OCLC) dedicated cataloging PCs.
+
+The 8-inch drive began to show up in 1971. Since they enabled developers and users to stop using the dreaded paper tape (which were easy to fold, spindle, and mutilate, not to mention to pirate) and the loathed IBM 5081 punch card. Everyone who had ever twisted a some tape or—the horror!—dropped a deck of Hollerith cards was happy to adopt 8-inch drives.
+{{./2.gif}}
+//Before floppy drives, we often had to enter data using punch cards.//
+
+Besides, the early single-sided 8-inch floppy could hold the data of up to 3,000 punch cards, or __80K__ to you. I know that's nothing today — this article uses up 66K with the text alone – but then it was a big deal.
+
+Some early model microcomputers, such as the Xerox 820 and Xerox Alto, used 8-inch drives, but these first generation floppies never broke through to the larger consumer market. That honor would go to the next generation of the floppy: the 5.25 inch model.
+
+By 1972, Shugart had left IBM and founded his own company, Shugart Associates. In 1975, Wang, which at the time owned the then big-time dedicated word processor market, approached Shugart about creating a computer that would fit on top of a desk. To do that, Wang needed a smaller, cheaper floppy disk.
+
+According to Don Massaro (PDF link), another IBMer who followed Shugart to the new business, Wang’s founder Charles Wang said, “I want to come out with a much lower-end word processor. It has to be much lower cost and I can't afford to pay you $200 for your 8" floppy; I need a $100 floppy.”
+
+So, Shugart and company started working on it. According to Massaro, “We designed the 5 1/4" floppy drive in terms of the overall design, what it should look like, in a car driving up to Herkimer, New York to visit Mohawk Data Systems.” The design team stopped at a stationery store to buy cardboard while trying to figure out **what size** the diskette should be. “It's real simple, the reason why it was 5¼,” he said. “5 1/4 was the smallest diskette that you could make that would not __fit in your pocket__. We didn't want to put it in a pocket because we didn't want it bent, okay?”
+
+Shugart also designed the diskette to be that size because an analysis of the cassette tape drives and their bays in microcomputers showed that a 5.25” drive was as big as you could fit into the PCs of the day.
+
+According to another story from Jimmy Adkisson, a Shugart engineer, “Jim Adkisson and Don Massaro were discussing the proposed drive's size with Wang. The trio just happened to be doing their discussing at a bar. An Wang motioned to a drink napkin and stated 'about that size' which happened to be 5 1/4-inches wide.”
+Wang wasn’t the most important element in the success of the 5.25-inch floppy. George Sollman, another Shugart engineer, took an early model of the 5.25” drive to a Home Brew Computer Club meeting. “The following Wednesday or so, Don came to my office and said, 'There's a bum in the lobby,’” Sollman says. “‘And, in marketing, you're in charge of cleaning up the lobby. Would you get the bum out of the lobby?’ So I went out to the lobby and this guy is sitting there with holes in both knees. He really needed a shower in a bad way but he had the most dark, intense eyes and he said, 'I've got this thing we can build.'”
+The bum's name was Steve Jobs and the “thing” was the Apple II.
+Apple had also used cassette drives for its first computers. Jobs knew his computers also needed a smaller, cheaper, and better portable data storage system. In late 1977, the Apple II was made available with optional 5.25” floppy drives manufactured by Shugart. One drive ordinarily held programs, while the other could be used to store your data. (Otherwise, you had to swap floppies back-and-forth when you needed to save a file.)
+{{./3.png}}
+The PC that made floppy disks a success: 1977's Apple II
+The floppy disk seems so simple now, but it changed everything. As IBM's history of the floppy disk states, this was a big advance in user-friendliness. “But perhaps the greatest impact of the floppy wasn’t on individuals, but on the nature and structure of the IT industry. Up until the late 1970s, most software applications for tasks such as word processing and accounting were written by the personal computer owners themselves. But thanks to the floppy, companies could write programs, put them on the disks, and sell them through the mail or in stores. 'It made it possible to have a software industry,' says Lee Felsenstein, a pioneer of the PC industry who designed the Osborne 1, the first mass-produced portable computer. Before networks became widely available for PCs, people used floppies to share programs and data with each other—calling it the 'sneakernet.'”
+In short, it was the floppy disk that turned microcomputers into personal computers.
+{{./4.jpg}}
+ Which of these drives did you own?
+The success of the Apple II made the 5.25” drive the industry standard. The vast majority of CP/M-80 PCs, from the late 70s to early 80s, used this size floppy drive. When the first IBM PC arrived in 1981 you had your choice of one or two 160 kilobyte (K – yes, just one K) floppy drives.
+Throughout the early 80s, the floppy drive became the portable storage format. (Tape quickly was relegated to business back-ups.) At first, the floppy disk drives were only built with one read/write head, but another set of heads were quickly incorporated. This meant that when the IBM XT PC arrived in 1983, double-sided floppies could hold up to 360K of data.
+There were some bumps along the road to PC floppy drive compatibility. Some companies, such as DEC with its DEC Rainbow, introduced its own non-compatible 5.25” floppy drives. They were single-sided but with twice the density, and in 1983 a single box of 10 disks cost $45 – twice the price of the standard disks.
+In the end, though, market forces kept the various non-compatible disk formats from splitting the PC market into separate blocks. (How the data was stored was another issue, however. Data stored on a CP/M system was unreadable on a PC-DOS drive, for examples, so dedicated applications like Media Master promised to convert data from one format to another.)
+That left lots of room for innovation within the floppy drive mainstream. In 1984, IBM introduced the IBM Advanced Technology (AT) computer. This model came with a high-density 5.25-inch drive, which could handle disks that could up hold up to 1.2MB of data.
+A variety of other floppy drives and disk formats were tried. These included 2.0, 2.5, 2.8, 3.0, 3.25, and 4.0 inch formats. Most quickly died off, but one, the 3.5” size – introduced by Sony in 1980 – proved to be a winner.
+The 3.5 disk didn't really take off until 1982. Then, the Microfloppy Industry Committee approved a variation of the Sony design and the “new” 3.5” drive was quickly adopted by Apple for the Macintosh, by Commodore for the Amiga, and by Atari for its Atari ST PC. The mainstream PC market soon followed and by 1988, the more durable 3.5” disks outsold the 5.25” floppy disks. (During the transition, however, most of us configured our PCs to have both a 3.5” drive and a 5.25” drive, in addition to the by-now-ubiquitous hard disks. Still, most of us eventually ran into at least one situation in which we had a file on a 5.25” disk and no floppy drive to read it on.)
+{{./5.jpg}}
+ The one 3.5” diskette that everyone met at one time or another: An AOL install disk.
+The first 3.5” disks could only hold 720K. But they soon became popular because of the more convenient pocket-size format and their somewhat-sturdier construction (if you rolled an office chair over one of these, you had a chance that the data might survive). Another variation of the drive, using Modified Frequency Modulation (MFM) encoding, pushed 3.5” diskettes storage up to 1.44Mbs in IBM's PS/2 and Apple's Mac IIx computers in the mid to late 1980s.
+By then, though floppy drives would continue to evolve, other portable technologies began to surpass them.
+In 1991, Jobs introduced the extended-density (ED) 3.5” floppy on his NeXT computer line. These could hold up to 2.8MBs. But it wasn't enough. A variety of other portable formats that could store more data came along, such as magneto-optical drives and Iomega's Zip drive, and they started pushing floppies out of business.
+The real floppy killers, though, were read-writable CDs, DVDs, and, the final nail in the coffin: USB flash drives. Today, a 64GB flash drive can hold more data than every floppy disk I've ever owned all rolled together.
+Apple prospered the most from the floppy drive but ironically was the first to abandon it as read-writable CDs and DVDs took over. The 1998 iMac was the first consumer computer to ship without any floppy drive.
+However, the floppy drive took more than a decade to die. Sony, which at the end owned 70% of what was left of the market, announced in 2010 that it was stopping the manufacture of 3.5” diskettes.
+Today, you can still buy new 1.44MB floppy drives and floppy disks, but for the other formats you need to look to eBay or yard sales. If you really want a new 3.5” drive or disks, I'd get them sooner than later. Their day is almost done.
+But, as they disappear even from memory, we should strive to remember just how vitally important floppy disks were in their day. Without them, our current computer world simply could not exist. Before the Internet was open to the public, it was floppy disks that let us create and trade programs and files. They really were what put the personal in “personal computing.”
+See also:
+6 Products That Became Technology Roadkill
+Searching the Internet B.G. (Before Google)
+The Birth and Rise of Ethernet: A History
+7 Awesome Bits of Tech That Just Freakin’ Disappeared
+30 Years Later: an Unfair Comparison between an IBM PC and an Apple iPhone4 (Infographic)