| 石山's profile山石记PhotosBlogLists | Help |
|
|
September 15 google搜索原理
from:http://dev.csdn.net/author/baggio785/f00f5da89d704c148bba798c1af0f6c2.html 这篇文章中,我们介绍了google,它是一个大型的搜索引擎(of a large-scale search engine)的原型,搜索引擎在超文本中应用广泛。Google的设计能够高效地抓网页并建立索引,它的查询结果比其它现有系统都高明。这个原型的全文和超连接的数据库至少包含24'000'000个网页。我们可以从http://google.stanford.edu/ 下载。 设计搜索引擎是一项富有挑战性的工作。搜索引擎为上亿个网页建立索引,其中包含大量迥然不同的词汇。而且每天要回答成千上万个查询。在网络中,尽管大型搜索引擎非常重要,但是学术界却很少研究它。此外由于技术的快速发展和网页的大量增加,现在建立一个搜索引擎和三年前完全不同。 本文详细介绍了我们的大型搜索引擎,据我们所知,在公开发表的论文中,这是第一篇描述地如此详细。除了把传统数据搜索技术应用到如此大量级网页中所遇到的问题,还有许多新的技术挑战,包括应用超文本中的附加信息改进搜索结果。 本文将解决这个问题,描述如何运用超文本中的附加信息,建立一个大型实用系统。任何人都可以在网上随意发布信息,如何有效地处理这些无组织的超文本集合,也是本文要关注的问题。 关键词 World Wide Web,搜索引擎,信息检索,PageRank, Google 1 绪论 1.1网络搜索引擎—升级换代(scaling up): 1.2 Google: 1. 3设计目标 1.3.2搜索引擎的学术研究随着时间的流逝,除了发展迅速,Web越来越商业化。 2. 系统特点 2.1 PageRank:给网页排序 Web的引用(链接)图是重要的资源,却被当今的搜索引擎很大程度上忽视了。我们建立了一个包含518‘000'000个超链接的图,它是一个具有重要意义的样本。这些图能够快速地计算网页的PageRank值,它是一个客观的标准,较好的符合人们心目中对一个网页重要程度的评价,建立的基础是通过引用判断重要性。因此在web中,PageRank能够优化关键词查询的结果。对于大多数的主题,在网页标题查询中用PageRank优化简单文本匹配,我们得到了令人惊叹的结果(从google.stanford.edu可以得到演示)。对于Google主系统中的全文搜索,PageRank也帮了不少忙。 2.1.1计算PageRank 文献检索中的引用理论用到Web中,引用网页的链接数,一定程度上反映了该网页的重要性和质量。PageRank发展了这种思想,网页间的链接是不平等的。 PageRank定义如下:我们假设T1…Tn指向网页A(例如,被引用)。参数d是制动因子,使结果在0,1之间。通常d等于0.85。在下一节将详细介绍d。C(A)定义为网页A指向其它网页的链接数,网页A的PageRank值由下式给出: PR(A) = (1-d) + d (PR(T1)/C(T1) + ... + PR(Tn)/C(Tn)) 注意PageRank的形式,分布到各个网页中,因此所有网页的PageRank和是1。 PageRank或PR(A)可以用简单的迭代算法计算,相应规格化Web链接矩阵的主特征向量。中等规模的网站计算26‘000'000网页的PageRank值要花费几小时。还有一些技术细节超出了本文论述的范围。 2.1.2直觉判断 PageRank被看作用户行为的模型。我们假设网上冲浪是随机的,不断点击链接,从不返回,最终烦了,另外随机选一个网页重新开始冲浪。随机访问一个网页的可能性就是它的PageRank值。制动因子d是随机访问一个网页烦了的可能性,随机另选一个网页。对单个网页或一组网页,一个重要的变量加入到制动因子d中。这允许个人可以故意地误导系统,以得到较高的PageRank值。我们还有其它的PageRank算法,见98页。 另外的直觉判断是一个网页有很多网页指向它,或者一些PageRank值高的网页指向它,则这个网页很重要。直觉地,在Web中,一个网页被很多网页引用,那么这个网页值得一看。一个网页被象Yahoo这样重要的主页引用即使一次,也值得一看。如果一个网页的质量不高,或者是死链接,象Yahoo这样的主页不会链向它。PageRank处理了这两方面因素,并通过网络链接递归地传递。 2.2链接描述文字(Anchor Text) 2.3其它特点除了PageRank和应用链接描述文字外,Google还有一些其它特点。 3有关工作 Web检索研究的历史简短。 3.1信息检索信息检索系统诞生在几年前,并发展迅速。然而大多数信息检索系统研究的对象是小规模的单一的有组织结构的集合,例如科学论文集,或相关主题的新闻故事。实际上,信息检索的主要基准,the Text Retrieval Conference(),用小规模的、有组织结构的集合作为它们的基准。大型文集基准只有20GB,相比之下,我们抓到的24000000个网页占147GB。在TREC上工作良好的系统,在Web上却不一定产生好的结果。例如,标准向量空间模型企图返回和查询请求最相近的文档,把查询请求和文档都看作由出现在它们中的词汇组成的向量。在Web环境下,这种策略常常返回非常短的文档,这些文档往往是查询词再加几个字。例如,查询“Bill Clinton”,返回的网页只包含“Bill Clinton Sucks”,这是我们从一个主要搜索引擎中看到的。网络上有些争议,用户应该更准确地表达他们想查询什么,在他们的查询请求中用更多的词。我们强烈反对这种观点。如果用户提出象“Bill Clinton”这样的查询请求,应该得到理想的查询结果,因为这个主题有许多高质量的信息。象所给的例子,我们认为信息检索标准需要发展,以便有效地处理Web数据。 3.2有组织结构的集合(Well Controlled Collections)与Web的不同点 Web是完全无组织的异构的大量文档的集合。Web中的文档无论内在信息还是隐含信息都存在大量的异构性。例如,文档内部就用了不同的语言(既有人类语言又有程序),词汇(email地址,链接,邮政编码,电话号码,产品号),类型(文本,HTML,PDF,图像,声音),有些甚至是机器创建的文件(log文件,或数据库的输出)。可以从文档中推断出来,但并不包含在文档中的信息称为隐含信息。隐含信息包括来源的信誉,更新频率,质量,访问量和引用。不但隐含信息的可能来源各种各样,而且被检测的信息也大不相同,相差可达好几个数量级。例如,一个重要主页的使用量,象Yahoo 每天浏览数达到上百万次,于此相比无名的历史文章可能十年才被访问一次。很明显,搜索引擎对这两类信息的处理是不同的。 Web与有组织结构集合之间的另外一个明显区别是,事实上,向Web上传信息没有任何限制。灵活利用这点可以发布任何对搜索引擎影响重大的信息,使路由阻塞,加上为牟利故意操纵搜索引擎,这些已经成为一个严重的问题。这些问题还没有被传统的封闭的信息检索系统所提出来。它关心的是元数据的努力,这在Web搜索引擎中却不适用,因为网页中的任何文本都不会向用户声称企图操纵搜索引擎。甚至有些公司为牟利专门操纵搜索引擎。 4 系统分析(System Anatomy) 4.1Google体系结构概述 4.2主要数据结构经过优化的Google数据结构,能够用较小的代价抓取大量文档,建立索引和查询。 4.2.1 4.2.2知识库 知识库包含每个网页的全部HTML。每个网页用zlib(见RFC1950)压缩。压缩技术的选择既要考虑速度又要考虑压缩率。我们选择zlib的速度而不是压缩率很高的bzip。知识库用bzip的压缩率接近4:1。而用zlib的压缩率是3:1。文档一个挨着一个的存储在知识库中,前缀是docID,长度,URL,见图2。访问知识库不需要其它的数据结构。这有助于数据一致性和升级。用其它数据结构重构系统,我们只需要修改知识库和crawler错误列表文件。 4.2.3文件索引 4.2.4词典 4.2.5 hit list 4.2.6正向索引实际上,正向索引已经部分排序。它被存在一定数量的barrel中(我们用64个barrels)。每个barrel装着一定范围的wordID。如果一篇文档中的词落到某个barrel,它的docID将被记录到这个barrel中,紧跟着那些词(文档中所有的词汇,还是落入该barrel中的词汇)对应的hitlist。这种模式需要稍多些的存储空间,因为一个docID被用多次,但是它节省了桶数和时间,最后排序器进行索引时降低编码的复杂度。更进一步的措施是,我们不是存储docID本身,而是存储相对于该桶最小的docID的差。用这种方法,未排序的barrel的docID只需24位,省下8位记录hitlist长。 4.2.7反向索引除了反向索引由sorter加工处理之外,它和正向索引包含相同的桶。对每个有效的docID,字典包含一个指向该词所在桶的指针。它指向由docID和它的相应hitlist组成的doclish,这个doclist代表了所有包含该词的文档。 doclist中docID的顺序是一个重要的问题。最简单的解决办法是用doclish排序。这种方法合并多个词时很快。另一个可选方案是用文档中该词出现的次数排序。这种方法回答单词查询,所用时间微不足道。当多词查询时几乎是从头开始。并且当用其它Rank算法改进索引时,非常困难。我们综合了这两种方法,建立两组反向索引barrel,一组barrels的hitlist只包含标题和anchor hit,另一组barrel包含全部的hitlist。我们首先查第一组索引桶,看有没有匹配的项,然后查较大的那组桶。 4.3抓网页 4.4Web索引分析 4.5搜索搜索的目标是提供有效的高质量的搜索结果。多数大型商业搜索引擎好像在效率方面花费了很大力气。因此我们的研究以搜索质量为重点,相信我们的解决方案也可以用到那些商业系统中。 4.5.1 Ranking系统 Google比典型搜索引擎保存了更多的web信息。每个hitlish包括位置,字号,大小写。另外,我们还考虑了链接描述文字。Rank综合所有这些信息是困难的。ranking函数设计依据是没有某个因素对rank影响重大。首先,考虑最简单的情况—单个词查询。为了单个词查询中一个文档的rank,Goole在文档的hitlist中查找该词。Google认为每个hit是几种不同类型(标题,链接描述文字anchor,URL,普通大字号文本,普通小字号文本,……)之一,每种有它自己的类型权重。类型权重建立了一个类型索引向量。Google计算hitlist中每种hit的数量。然后每个hit数转换成count-weight。Count-weight开始随hit数线性增加,很快逐渐停止,以至于hit数与此不相关。我们计算count-weight向量和type-weight向量的标量积作为文档的IR值。最后IR值结合PageRank作为文档的最后rank 对于多词查询,更复杂些。现在,多词hitlist必须同时扫描,以便关键词出现在同一文档中的权重比分别出现时高。相邻词的hit一起匹配。对每个匹配hit 的集合计算相邻度。相邻度基于hit在文档中的距离,分成10个不同的bin值,范围从短语匹配到根本不相关。不仅计算每类hit数,而且要计算每种类型的相邻度,每个类型相似度对,有一个类型相邻度权type-prox-weight。Count转换成count-weight,计算count-weight type-proc-weight的标量积作为IR值。应用某种debug mode所有这些数和矩阵与查询结果一起显示出来。这些显示有助于改进rank系统。 4.5.2反馈 5.1 5.2系统执行搜索引擎抓网页和建立索引的效率非常重要。Google的主要操作是抓网页,索引,排序。很难测试抓全部网页需要多少时间,因为磁盘满了,域名服务器崩溃,或者其它问题导致系统停止。总的来说,大约需要9天时间下载26000000网页(包括错误)。然而,一旦系统运行顺利,速度非常快,下载最后11000000网页只需要63小时,平均每天4000000网页,每秒48.5个网页。索引器和网络爬行机器人同步运行。索引器比网络爬行机器人快。因为我们花费了大量时间优化索引器,使它不是瓶颈。这些优化包括批量更新文档索引,本地磁盘数据结构的安排。索引器每秒处理54个网页。排序器完全并行,用4台机器,排序的整个过程大概需要24小时。 5.3搜索执行改进搜索执行不是我们研究的重点。 6.1未来的工作大型Web搜索引擎是个复杂的系统,还有很多事情要做。 6.2高质量搜索当今Web搜索引擎用户所面临的最大问题是搜索结果的质量。 6.3可升级的体系结构除了搜索质量,Google设计成可升级的。 6.4研究工具 图2 Google系统的工作流程图 ①Google使用高速的分布式爬行器(Crawler)系统中的漫游遍历器(Googlebot)定时地遍历网页,将遍历到的网页送到存储服务器(Store Server)中。
图3 顺排档索引和Hit的存储结构 值得注意的是,当特殊索引项来自Anchor Text时,特殊索引项用来表示位置的信息(8位)将分为两部分:4位表示Anchor Text出现的具体位置,另4位则用来与表示Anchor Text所链接网页的docID相连接,这个docID是由URL Resolver经过转化存入顺排档索引的。 图4 倒排档索引结构
发表于 @ 2005年12月13日 12:04 PM | 评论 (10)
网站应具有清晰的层次结构和文本链接。每个网页应至少可以通过一个静态文本链接打开。 由于大部分搜索引擎 Spider 查看网站的方式与文本浏览器一样,所以可使用诸如 Lynx 的文本浏览器来检查您的网站。如果因采用了 Javascript、Cookie、Session ID's、frames (框架)、DHTML 或 Flash 等复杂技术,而导致在文本浏览器中无法看到网站的所有网页,则搜索引擎 Spider 在收录您的网站时可能会遇到麻烦。 一旦您的网站在线后,请将其提交给 Google,具体网址为: http://www.google.com/addurl.html 。 网页应面向用户,而不是面向搜索引擎。对用户应保持诚实无欺的态度,提交给搜索引擎的内容应与显示给用户的内容一致。 请不要使用隐藏文本或隐藏链接。 如果您认为有网站违背了 Google 质量指南,请告诉我们,网址为: http://www.google.com/contact/spamreport.html 。对于网络作弊行为,Google 希望能开发出易于推广的自动解决方案,而尽量减少人为的干预。我们将利用收到的作弊行为报告来创建广泛适用的算法,从而识别并阻止未来的作弊行为。 July 06 内存调试技巧
October 26 Active Movies动感英语003_050103.rmvb
Action moive can't get over 无法忘记...,从...阴影走出来 dwell on 留守、驻扎 We shouldn't dwell on someone's past Get down 蹲下、趴下 Freeze 不许动 Duck 把头低下 the hell 强调,有语气 What (the hell) is going on? (究竟)出了什么事情? have the right to do 有权干... You have right to keep slient. Moon river
动感英语004_050104.rmvb
Preview 预告片 play chicken 叫板
Don't be a baby = don't complain 不要抱怨 he is such a baby 他是一个喋喋不休的人 keep my edge 保持最好状态 I have to speak english Just in order to keep my edge loose my edge 失去最好状态 hone your skills 提升你...技能 California Deaming
动感英语005_050105.rmvb
Screen Adaption 改编 take your pick 随你挑
That's a bad line He is kind o fa ladies man 他很有女人缘 such a ladies man 如此又女人缘 I don't buy = I don't believe 动感英语006_050106.rmvb
Log keeper 场记 It's up to you 由你来决定
what do you mean?你什么意思? Can I get a minute 等我一下,好吗?(注打电话应该用just a minute) what's matter (with you)? =
what's got into you? = what's going on? propose to sb.向某人求婚
get engaged to 订婚 A got engaged to B got married to 结婚 A got married to B Take my breath away
动感英语007_050107.rmvb
Crew 剧组 bust sb. ass 难为某人、努力干
smooth 对...很有一套(He is very smooth with...) 形容一个人处理事情很厉害(he is a smooth) 形容一个很有说服力、感染力(he has a smooth manner) sneak 潜入(Move sliently without being seen)
head to somewhere(= go somewhere)
I am heading home(here there)(后面跟home here there时没有to) memorize = to remember something记下
I have to memorize forty vocabulary today 动感英语010_050110.rmvb
thriller 惊粟片 bluff = lie 欺骗、骗人、撒谎
(have)get sth. under control(使某事情在控制之下)
Everything is under control get sth. out of control 失控 My weight is out of control any idea...(接一个句子)= do you know
Any idea when the mall will be closed? Any idea why Li is not here? <<Circle of Life>>
动感英语013_050113.rmvb
Western 西部片 straight up 完全正确(非正式场合)
= absolutely = exactly = right = correct = You said it = you can say that again(正式场合) gorgeous = very beautiful
pretty < beautiful < gorgeous(程度越来越高) cool: that girl is so cool killer: She has killer legs(她的腿太美了) He is a lady killer(他是万人谜) hot:性感的 personality有个性的 nice:形容人很好,形容东西好 You could do better 你可以做的更好
feed 喂 You have to feed your dreams 你要让你的梦想成长起来(feed很逼真)
《Can You Fell the Love Tonight》
动感英语017_050117.rmvb
Sound track 原声碟 a head start(on sb.) 走在别人的前面(有点笨鸟先飞的意思)
You should get up early and study Get a head start on your classmates He is a slow runner, but he get a head start on the others so he might win the race I have no idea that(语气强)... = I don't know that...
mad at sb. 生某人气 You'd better go talk to Frank He is really mad at you What's up? = 熟人之间问候(How are you?) What't up? = 什么事? <<As Time Goes by>>
动感英语018_050118.rmvb
costumer 服装师 costume 服装 wardrobe 行头 What's up Hey, What's up 打招呼
出什么事了?= What's matter with you? make a difference
My friend Joe gives money to hte poor because he really wanna make a difference in people's lives(使穷人的生活好起来) Your heart is in it 表示你愿意干这个事情
You're up = it's your turn该轮到你了
I just tried the spicy fish Now you're up <<Ain't No Mountain high Enough>>
动感英语019_050119.rmvb
stunt 替身 stuntman stuntwoman pawn sth. off on sb.= sell sth. cheaply 把某东西廉价卖给谁(把某人推给某人)
pawn her off on me go for对...感兴趣 You can't pawn your old car off on me I didn't go for it verge = edge 边缘
on the verge of... I am really angry at you I am on the verge of hitting you
我真的生气了,我要打你 hold someone back 拖某人后腿
I'm doing well at work but the fact that I can't speak English is holding my back I never get promoted 我工作不错,但是我不会讲英语的事情拖了我的后腿,从来没有得到提升 《My God》
动感英语020_050120.rmvb
director 导演 tiger 爱称,American fathers often call their sons tiger(叫自己儿子)
I'm so proud of your, tiger Go, help your mom. tiger pumpkin 与tiger对应,是指女儿
honey = sweet heart = darling 泛泛指,不分男女 strip (to strip = to take off your clothes)
The doctor said to go into the exam room and strip () Strip the shirt off it's a terrible color on you go over sb. heads (越过...向更高级别告状)
《Gravy》
Gimme = give me mashed potatoes 土豆泥 动感英语021_050121.rmvb screenplay 电影剧本 screenplay writer 编剧 April 11 A great website about project management(key practices)Table of Contents
refering to http://aelinik.free.fr/cmm/default.htm March 30 Pro ASP.NET 2.0---Learning notesPART 1 Core Concepts
■CHAPTER 1 Introducing ASP.NET 1.1 Some of the differences between ASP.NET and earlier web development platforms include the following:
• ASP.NET features a completely object-oriented programming model, which includes an event-driven, control-based architecture that encourages code encapsulation and code reuse. • ASP.NET gives you the ability to code in any supported .NET language (including Visual Basic, C#, J#, and many other languages that have third-party compilers). • ASP.NET is also a platform for building web services, which are reusable units of code that other applications can call across platform and computer boundaries. You can use a web service to do everything from web-enabling a desktop application to sharing data with a Java client running on Unix. • ASP.NET is dedicated to high performance. ASP.NET pages and components are compiled on demand instead of being interpreted every time they’re used. ASP.NET also includes a finetuned data access model and flexible data caching to further boost performance. These are only a few of the features, which include enhanced state management, practical data binding, dynamic graphics, and a robust security model. 1.2 Seven Important Facts About ASP.NET
If you’re new to ASP.NET (or you just want to review a few fundamentals), you’ll be interested in the following sections. They introduce seven touchstones of .NET development. Fact 1: ASP.NET Is Integrated with the .NET Framework Fact 2: ASP.NET Is Compiled,Not Interpreted Fact 3: ASP.NET Is Multilanguage
Fact 4: ASP.NET Runs Inside the Common Language Runtime
Fact 5: ASP.NET Is Object-Oriented
Fact 6: ASP.NET Is Multidevice and Multibrowser
Fact 7: ASP.NET Is Easy to Deploy and Configure ■CHAPTER 2 Visual Studio 2005
Here are the four most signifficant changes comparing to Visual Studio 2003: • Projectless development:
Visual Studio no longer clutters your web projects with extra development files (such as .csproj and .sln). One obvious benefit of this model is that you can deploy exactly what you develop, without needing to filter out just a subset of the files. However, as you’ll see in this chapter, the concept of projectless development is slightly overstated. Visual Studio still stores some information in a solution file (such as breakpoints and build settings), and it quietly stows that file away under a user-specific directory. However, there’s a significant difference—these hidden solution files aren’t required. Essential details (such as project references) are stored right in the web.config file. You’ll learn about projectless development in the “Websites in Visual Studio” section of this chapter. 23
• New compilation model: Visual Studio is no longer responsible for compiling your code. Instead, ASP.NET takes on that responsibility exclusively. This gives Visual Studio more flexible debugging, and it simplifies deployment on different platforms (for example, 32-bit and 64-bit Windows). It also allows you to combine web pages written in C# with web pages written in another .NET language (such as Visual Basic) in the same project.
• New code model:
The shift in the compilation model also reduces the differences between the code-behind model and the code-inline model of writing web pages, both of which Visual Studio now supports. However, the syntax for code-behind is subtly different from that used for Visual Studio 2003 web pages, and you’ll need to perform a one-way conversion operation to edit your web application in Visual Studio 2005. You’ll learn about the coding model in “The Coding Model” later in this chapter.
• Integrated test web server: If you’ve programmed with Web Matrix (a scaled-down design tool used with ASP.NET 1.x), you’ll recognize the new integrated web server, which allows you to run your web pages without setting up virtual directories or deploying your website.
listing some of file types in ASP.NET web applications
Ends with .aspx: These are ASP.NET web pages (the .NET equivalent of the .asp file in an ASP application). They contain the user interface and, optionally, the underlying application code. Users request or navigate directly to one of these pages to start your web application. Ends with .ascx: These are ASP.NET user controls. User controls are similar to web pages, except that they can’t be accessed directly. Instead, they must be hosted inside an ASP.NET web page. User controls allow you to develop an important piece of the user interface and reuse it in as many web forms as you want without repetitive code. Ends with .asmx: These are ASP.NET web services. Web services work differently than web pages, but they still share the same application resources, configuration settings, and memory. web.config: This is the XML-based configuration file for your ASP.NET application. It includes settings for customizing security, state management, memory management, and much more. Visual Studio adds a web.config file when you need it. (For example, it adds a web.config file that supports debugging if you attempt to run your web application.) When you first create a website, you won’t have a web.config file. global.asax: This is the global application file. You can use this file to define global variables and react to global events, such as when a web application first starts (see Chapter 5 for a detailed discussion). Visual Studio doesn’t create a global.asax file by default—you need to add it if it’s appropriate. Ends with .cs: These are code-behind files that contain C# code. They allow you to separate the application from the user interface of a web page. The code-behind model is introduced in this chapter and used extensively in this book. CHAPTER 2 ■ VISUAL STUDIO 2005 24
■CHAPTER 3 Web Forms ■CHAPTER 4 Server Controls ■CHAPTER 5 ASP.NET Applications ■CHAPTER 6 State Management March 07 有关软件开发一些文档从今天开始要写一个软件项目的需求分析文档,原来写过几次,不过具体格式忘了,只是记得是国标那个模板。打算到网上找个模板,发现了一个网站 very good, to share with you.
March 06 PMBOK learning notesSection1 The Project Management Framework
chapter1 Introduction
1.1what is project?
Project is a temporary endeavor undertaken to create a unique product, service or result.
1.1.1project characteristics
1.Temporary
Temporary means that a project has a definite beginning and a definite end.
In addition, Temporary does not generally apply to the products, services or result created by the project. Most projects are undertaken to a create a lasting outcome. For example, to elect a national monument.
The temporary nature of the porjects may apply to other aspects of the endeaor as well:
a. the projec has a limited time frame
b. th project team is disbanded and the team members reassigned when the project ends
2.Unique Products, Service, or Results
A project creates unique deliverables, which are products, service or results.
3.Progressive Elaboration Progressive elaboration means developing in steps, continuing by increments 1.1.2 Projects vs. Operational Work They share many of the following characterists: Performed by people Constrained by limited resources Planned, executed, and controlled
Projects and operations differ primarily in that operations are ongoing and repetitive, while projects are temporary and unique 1.1.3 Project vs. Strategic Planning Projects are often utilized as a means of achieving an orgnazation's strategic plan. 1.2 What is a project management? Project management is the application of knowledge, skills, tools techniques to project activities to meet project requirements. Project mamagement is accomplished through the application and integration of the project management processes of initiating, planning, executing, monitoring and controlling, and closing. The project manager is the person responsible for accomplishing the project objectives. (continue....) |
|
||||||||||||||||||||||||||
|
|