How do western news organizations get such great quotes from their Chinese sources? Linghu Lei, creative director at New Weekly and founder of the online magazine Rocky Week (home of The iPod Times), describes a telephone interview he had with a New York Times reporter:

So this is how the New York Times conducts an interview

Unexpectedly, I received an interview call from the New York Times today. The theme was "How to you perceive the entrance of Vogue and other western fashion magazines into China?"

Reading the New York Times in the past, I had always felt that their reports were not only outstanding, but that the material and the reaction of the interviewee worked together to serve the content of the article - it's not easy to achieve this. When this kind of narrative + interview is written in the domestic media, you get answers to questions you haven't asked, or the answers the interviewee gives are not really in accordance the perspective and outlook you had planned for the article.

During the interview process today, the reporter from the New York Times first spoke a good deal about why he wanted to do the interview, and then after I had answered the first question, he said that he "identified" with my viewpoint and felt that it was "special." Then he said, "Your viewpoint is extremely close to what I want to communicate with in article," and wanted me to discuss it further. To express this identification that he felt toward my viewpoint, I elaborated on the point I had just mentioned in passing, and then expanded a bit on his viewpoint. Then he echoed my viewpoint. During this process of communication, the New York Times reporter always led me to understand that what I said was precisely what he needed for his article. My viewpoint supported his argument exactly. So unwittingly I became his news spokesperson. At the end, for the final question, the reporter summarized our entire "discussion" and asked me if I agreed - at this point, there was no way I could deny what I had just said, so - right, "that's correct."

The reporter seemed satisfied that he had obtained all of the viewpoints he needed for "the article he wanted to write," and he said "Thank you" and hung up. This, then, is a New York Times interview.

--written 12 September

See this and related posts at Language Log for a further analysis of the phenomenon.

The cover of Beijing Times features a picture of a basketball match between China and Saudi Arabia.

2. China's 'second-string' team overwhelms Saudi Arabia 98-10 in the semi-final of the Asian Men's Basketball Championships;

3. Insurgents launched at least a dozen suicide bombings that ripped through Baghdad in rapid succession on Wednesday, killing almost 150 people in a coordinated assault;

4. Three prime culprits who kidnapped Wu Ruopu, a well-known actor executed (for more about the kidnapping, see the botttom of this
Danwei post from 2004));

5. An accident on Beijing's North Forth Ring Road caused a six-hour-long traffic jam.

Hot news this week: Google has launched a blog search tool. I think we were always expecting a blog: operator for the regular Google search site, which I think I would have liked better. But Google decided to come up with a separate engine.

昨天Google发布了BlogSearch
正确的提交自己的RSS还是要通过SiteMaps：

今天重新用了一下Google的SiteMaps网站：使用了一下提交网站认领（确认）功能。

具体信息如下：Googlebot要求你在网站的FEED目录下创建一个空文件名，比如：

确认网站： http://www.niernier.com/
Google 已经抓取了您的网站 http://www.niernier.com/ ，并将向您展示某些结果。然而，为确保我们不会将您网站的隐私信息透露给他人，我们需要确认您的所有权。

« 返回我的 Sitemaps

在网站上touch这个文件后：GOOGLE1cd1cd73cbe423a2.html
点击“检查”Googlebot会立刻访问这个文件：记录到的日志如下
66.249.65.141 - - [15/Sep/2005:16:13:27 +0800] "HEAD /GOOGLE1cd1cd73cbe423a2.html HTTP/1.1" 200 0 "-" "Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)"

Googlebot真是够省的，连内容都不要，只要HEAD发现这个文件存在即可。
如果成功：这个网站是你的了！网站被认领的好处在于你可以看到Googlebot的访问日志：从而帮助你的网站进行死链检查。

比如我的网站的“ 统计信息”如下：

网站： http://www.chedong.com/blog/
在我们的常规抓取过程中发现的网址

网址 错误
http://www.chedong.com/blog/archives/000019.html HTTP 错误
http://www.chedong.com/blog/archives/000020.html HTTP 错误
http://www.chedong.com/blog/archives/000027.html HTTP 错误

Che Dong posted a photo:

昨天Google发布了Blog搜索：
google.com/blogsearch
如果发现自己的blog查不到，到Google的SiteMaps： www.google.com/webmasters/sitemaps/
上去检查一下：没错SiteMaps就是一个RSS的提交入口。

我发现我提交的RSS都能查的到，而查不到的 www.adayang.com 是Google同步失败的FEED，并非WordPress发布系统的问题。

其实Blog搜索的关键还是RSS FEED的发现和同步，和传统的基于SPIDER的网页抓取分析相比，基于RSS的搜索引擎成本优势明显：
1 同步速度快：因为只需要同步RSS即可，减少了对网站的大量抓取，我看到很多RSS甚至是小时级别的同步抓取更新（从apache日志里也可以看到）；
2 内容直达：通过RSS进行内容索引，减少了对索引页，专题目录页的抓取，便于用户直接找到具体的内容而不是内容入口页；
3 内容提取简单：因为RSS将标题/摘要已经提供了，避免了后台进行内容提取的工作，可以轻松的将和主题无关的导航，页头/页尾和广告过滤掉。

Hu Bin in a traditional Chinese mood

The four-image series is titled Another set of alternative Hu Bin portraits: plums, orchids, bamboo and chrysanthemum added to the naked body strike the eyes. The image pictured here shows Hu Bin posing with plum blossoms.

Now this year, we find the letters sections of newspapers full of people discussing...the morality of college women having nude photoshoots done. Was last year's flap unresolved? Are colleges still shirking their responsibilities in loco parentis? Or should we simply be thankful that it's nothing worse this year, like the news that 20-30% of college women moonlight as working girls? (oh, wait, that idea was floated back in 2003 as well.)

A selection of letters printed in Beijing's Legal Mirror provides a nice cross-section of opinions - readers range from thinking that nude photography is unethical and unhealthy to simply worring that the young women will be taken advantage of (curiously, there has been no discussion of men doing these photo shoots...).

Gao Fusheng: Currently, the country's laws and regulations have a "blind spot" in regard to oversight of this new practice of "fully nude photography," and there is no clear line marking the "degree of exposure." That line is in the hands of the photographer doing the shoot. While there exists an unwritten rule in the profession that a male photographer must always have a female assistant present; i.e. the photographer must not come into contact with the nude subject at any time, so the female assistant must carry out the photographer's wishes; and the negatives are handed over to the subject at the same time as the photos. But are these "professional rules" effective?

If unlawful businessmen seize the opportunity to profit by copying and distributing large numbers of photograps, then what can be done? If these photos of "indecent exposure" are put to use by an unscrupulous minority, then the losers are those young college women themselves. Even if they try to protect their rights through legal means after the fact, getting proof is no easy matter!

Xu Linlin: I cannot readily agree with many of the misgivings people have. First, recognizing the beauty of a youthful body is an expression of self-confidence in one's life. The self-discovery of these young women is good, not harmful. Second, voluntarily having nude pictures taken has no connection with morality or the law. Third, as adult women, these college students surely know that nude photography is a closely private matter, and they know how to protect themselves against the unexpected.

But in my view, having nude photos taken is far from being a "symbol of fashion." College administrations and the parents of these women should stay on their toes, to prevent these young women from being unwittingly led astray and bringing trouble onto themselves.

Yin Jianguang: Having nude photos taken is without a doubt a precursor to having one's rights infringed. And if the photos are let out, there is no way to repair the harm done to you even if you pursue legal compensation. Besides, this is an unhealthy outlook on consumer fashion. Every penny that students spend was earned through their parents hard labor, so they shouldn't waste it in this manner.

This kind of "popularity" reflects the fact that educational activities are not making a deep impression in the minds of students. Behind this fashion is an expression of a sense of spiritual beauty that is left unsatisfied.

最近习惯先写主题，然后再慢慢完善：）

单页面应用程序（Single Page Application）是只有一个页面Web应用程序。WikiPedia上这样描述的：

A single page application (SPA) is a web application that runs entirely in the client web browser, typically using a combination of HTML, JavaScript, and CSS. The application modifies the web page's own data structures through its DOM tree, and makes its changes persistent when the user invokes the browser's Save Page command to save the (modified) web page to disk.

举几个例子吧：

• Solitaire Mahjongg - （麻将对对碰）非常Cool!
• Time-Tracker - 简单的时间管理程序
• TiddlyWiki - 更好玩的Wiki

作者：Andy（http://blog.xintiantang.com/andy），版权所有，转载请先联系我

这篇文章于2005/08/30 20:00开始，计划在一个月内完成，所以将会人工置顶。内容有点儿乱，先把想的和做的记下来，然后再慢慢整理。同时欢迎提出你的看法。更新记录：2005/08/30 20:00，2005/08/31 14:45，2005/08/31 21:21，2005/09/01 16:24，2005/09/01 21:53，2005/09/02 22:13，2005/09/14 13:10，2005/09/15 09:35

对于普通用户来说，使用传统的桌面应用程序（可能采用不同语言和工具开发的）可能存在这样那样的限制，比如界面风格千变万化、操作习惯不统一、学习难度大、无法或很难和外部应用通信等。而传统的服务器端应用程序或网站也存在许多限制，比如网络连通可能出现问题、响应速度慢、不能保证服务器端始终可用、担心数据安全和个人隐私问题、不够灵活、使用不方便等。

 而基于浏览器的桌面应用程序正是综合桌面应用程序和服务器端应用程序的功能产生的。它包含以下优点：

• 跨平台的、统一的、简单易用的操作界面（基于浏览器，比如IE或FireFox，用户只须要熟悉浏览器，而不必关心是机器运行的Windows、Mac OS X还是Linux，甚至不要OS，只要能运行Browser和JVM的硬件）
• 真正的去中心化应用（Decentralized application），每个应用即可以当成桌面应用程序，又可以当成服务器端应用程序（只要联网并开放端口就可以允许其他人远程运行）。这是Web 3.0吗？
• 应用XML进行通信，不用考虑本地服务（Local service）还是远程服务（Remote service）之间的区别
• 轻量级、易开发
• 。。。

基于浏览器的桌面应用程序（Browser-based desktop application）的定义：一个包含嵌入式网站服务器(Embedded web server)的能够在浏览器内的客户端的独立运行（Stand alone）的应用程序。

我把基于浏览器的桌面应用程序（Browser-based desktop application）简称为BBDA。其实BBDA和AJAX技术一样，也是一个buzzword，而不是什么新技术，只不过把几种通用的软件或技术组合在一起而已。和传统的重量级的Portal系统不同的是，BBDA是轻量级网络应用（Light weight web application），XML（Feed / RSS / XML-RPC / SOAP）是联系各个服务或应用的纽带。

其实，BBDA很符合目前热炒的WebOS（基于网络浏览器的操作系统）的特性，这方面Google做了个最好的Demo：Google Desktop, 在GD中，一切内容的表现都是网页（虽然隐藏了浏览器），并且很容易和Google网站进行通信（比如同时搜索本地和搜索Web，利用Feed（RSS）或XML-PRC或SOAP之类的协议读新闻、邮件、天气预报、股票行情等，微软该称RSS为Feed好像很有道理吧？）。现在GDS（Google Desktop Search）只是GD的一个模块而已，百度的桌面搜索只是单纯地模仿了GDS，却没有猜到原来Google是要做GD，好傻的百度。Google这几年的收购和挖人活动（Google Net, Google Browser等），看来一切都是为了WebOS而努力。这里顺便说一下Yahoo刚收购的Konfabulator，Konfabulator和GD有很大不同，我觉得它只是包含许多小软件或和网站交互的工具，只是个玩具箱而已，要做WebOS靠这个可以吗？看来Yahoo的眼光似乎始终不如Google。找到这篇文章：GoogleOS? YahooOS? MozillaOS? WebOS? 居然和我的想法如此相似（写这个之前我绝对没看过这篇文章）！

由于我主要做Java开发，所以只列出可能涉及的Java软件包：Browser(IE/FireFox), JDK, Tomcat/Jetty, HSQLDB/Apache Derby, Hibernate, Spring, Groovy, DOM4J,Apache Http Client, Lucene, Struts/WebWork/JSF/Tapestry/Wicket, DWR(AJAX), Apache Web Services(SOAP/XML-RPC), Apache POI(Word/Excel/ PDF), MP3 ID TAG 3 lib ,JPEG EXIF lib...

我的目的是组合几种通用的JAVA软件，用来进行快速的BBDA开发。对于开发者而言，只要选好合适的软件，并合理组合起来，完全可以达到或超过Ruby on rails的功能和开发效率。同时，对于普通用户而言，只要打开浏览器输入http://localhost就会进入Dashboard（ProtoPage是个很好的Prototype^_^，Start也不错），然后就可以管理日记（blog）、日历、通讯录、照片（flickr?）、音乐(itunes?)、文件、网摘(delicious？)、阅读订阅的RSS(bloglines?)、查看邮件（web mail）、新闻之类等，并可以进行搜索、共享、发布、与外部应用或网站同步等，构成一个完整的个人门户系统(Personal portal），XML（Feed / RSS / XML-RPC / SOAP）是联系各个服务或应用的纽带。

画了个草图（持续完善中）：

关于嵌入式网站服务器(Embedded web server)：

我这几天试用了Jetty和Tomcat。总以为Jetty很小呢，最适合做BBDA呢，结果发现软件包也很大了，另外很多功能不如Tomcat，Tomcat Embed 5.5由于使用Ecipse JDT，现在已经很小了（3M左右），而且标准，所以选了Tomcat Embed（这里下载）。下面的代码是Tomcat Embed的启动程序。

Start.java:

package com.terac.tomcat;
import java.net.InetAddress;
import java.net.UnknownHostException;
import org.apache.catalina.Context;
import org.apache.catalina.Engine;
import org.apache.catalina.Host;
import org.apache.catalina.LifecycleException;
import org.apache.catalina.connector.Connector;
import org.apache.catalina.startup.Embedded;

public class Start {
 public static void main(String[] args) {
  try {
   String path = System.getProperty("SERVER_HOME");
   if(path!=null&&!"".equals(path)){    
    int port = Integer.parseInt(System.getProperty("SERVER_PORT","8080"));
    startTomcat(path,port);
   }   
  } catch (Exception e) {
   e.printStackTrace();
  }
 }
 private static void startTomcat(String path,int port) throws UnknownHostException,
   LifecycleException {
  System.setProperty("catalina.home", path);
  Embedded embedded = new Embedded();
  Engine engine = embedded.createEngine();
  engine.setName("Engine");
  Host host = embedded.createHost("localhost", "web");
  Context rootContext = embedded.createContext("", "");
  rootContext.setPrivileged(true);
  host.addChild(rootContext);
  engine.addChild(host);
  engine.setDefaultHost("localhost");
  embedded.addEngine(engine);
  Connector connector = embedded.createConnector((InetAddress) null,port, false);
  embedded.addConnector(connector);
  embedded.start();
 }
}

start.bat:

@set classpath=./lib/所有的jar文件
@java -DSERVER_HOME=%cd% -DSERVER_PORT=80 -classpath %classpath% com.terac.tomcat.Start

关于嵌入式数据库(Embeded database)：

Apache Derby虽然比HSQLDB强大，但还是太大了，尤其是当用到IBM的JDBC Driver时。所以只好选HSQLDB。下面的代码是利用ServerContextListener在Web应用程序启动时启动HSQLDB服务器。

HSQLDBServletContextListener.java:

package com.terac.hsqldb;
import java.sql.Connection;
import java.sql.DriverManager;
import java.sql.Statement;
import javax.servlet.ServletContext;
import javax.servlet.ServletContextEvent;
import javax.servlet.ServletContextListener;

public class HSQLDBServletContextListener implements ServletContextListener {
 public void contextInitialized(ServletContextEvent evt) {
        try {
            ServletContext context = evt.getServletContext();
            String params = "port=9001;silent=true;trace=false;database.0=" +
                context.getRealPath("/") + "/WEB-INF/hsqldb/terac";
            org.hsqldb.Server server = new org.hsqldb.Server();
            server.putPropertiesFromString(params);
            server.setLogWriter(null);
            server.setErrWriter(null);
            server.start();
        } catch (Exception e) {
            System.err.println(e);
        }
    }
    public void contextDestroyed(ServletContextEvent evt) {
        Connection conn = null;
        try {
            Class.forName("org.hsqldb.jdbcDriver");
            conn = DriverManager.getConnection("jdbc:hsqldb:hsql://localhost:9001",
                    "sa", "123456");
            Statement stmt = conn.createStatement();
            stmt.executeUpdate("SHUTDOWN;");
        } catch (Exception e) {
            //
        }
    }
}

web.xml中需要加入listener：

com.terac.hsqldb.HSQLDBServletContextListener

关于全文检索引擎（Full text search engine）：

Apache Lucene应用广泛，也足够强大，高亮显示（Highlighting）需要自己实现，自动摘要（autosummarize）需要Classifier4j 。下面是常用类的封装代码：

Indexer.java:

还未整理好

Searcher.java:

还未整理好

关于Shell：

有些时候，用户不满足已经有的服务，需要更灵活的、或者定制的功能，一个简单的、易学的、解释性的、交互式的Shell就显得很必要，比如Windows提供的cmd或Linux里的bash。研究了一下，发现Groovy正好能满足需要，“for (ch in "Hello world!") {println ch}”，这个稍微复杂点儿的Hello world很好懂吧？

关于时间服务（Time Service）：

Windows里有计划任务（Schedule task），Linux里有Cron job，这个对于WebOS也是必须的，比如定时检查邮件，定时抓新闻等等都需要定时器（Timer），Spring framework封装了好几种Time service，足够了。

关于对象关系映射工具（OR mapping tools）：

Hibernate是最常用最强大的工具。下面是Hibernate Session的封装代码（也可以利用Spring framework来管理）：

HibernateUtil.java:

import net.sf.hibernate.HibernateException;
import net.sf.hibernate.SessionFactory;
import net.sf.hibernate.Session;
import net.sf.hibernate.cfg.Configuration;
import org.apache.log4j.Logger;

public class HibernateUtil {
 private static Log log = LogFactory.getLog(HibernateUtil.class); 
 public static final ThreadLocal threadSession = new ThreadLocal();
 private static final SessionFactory sessionFactory;
 static {
  try {
   sessionFactory = new Configuration().configure().buildSessionFactory();
  } catch (Throwable ex) {
   // We have to catch Throwable, otherwise we will miss
   // NoClassDefFoundError and other subclasses of Error
   log.error("Building SessionFactory failed.", ex);
   throw new ExceptionInInitializerError(ex);
  }
 }
    public static Session currentSession() throws HibernateException {       
        Session s = (Session) threadSession.get();
        if (s == null) {
            s = sessionFactory.openSession();
            threadSession.set(s);
        }
        return s;
    }
    public static void closeSession() throws HibernateException {
        Session s = (Session) session.get();
        session.set(null);
        if (s != null)
            s.close();
    }
}

关于模型-视图-控制框架（MVC framework）：

现在的MVC framework太多了，传统的有Struts/WebWork，新的有JSF/Tapestry/Wicket/Flex&Laszlo(Flash RIA) ，Wicket网站列出了55种之多！我试过JSF，始终不能另我满意，正在研究Wicket和RIFE(这两个都支持AJAX)，新加一个Stripes（采用Java 5 Annotations，所以不需要XML配置文件，非常简单）。

关于文件系统（File system）：

Apache Commons Virtual File System - Commons VFS provides a single API for accessing various different file systems. It presents a uniform view of the files from various different sources, such as the files on local disk, on an HTTP server, or inside a Zip archive. 越来越好玩了：）

另外，发现在Windows系统下PHP5也可以在Tomcat下运行（Running php 5.x on windows using tomcat 4.x or 5.x - Simple Thoughts - Java and Web Software）。BBDA with PHP：），不知道PHP的基于文本的数据库都有哪些？我试了一下，phpinfo和一个简单的基于文本的PHP留言本可以使用，phpMyAdmin和几个PHP的blog不能运行，Tomcat直接Crash。

我已经把BBDA框架，以前写的一个非常简单的Struts/Hibernate单用户相册程序，还有一个基于文本的PHP留言本做好打包了，下载后解压缩就能用。（运行PHP需要设Path，整个软件包稍后提供）

Looks like Word, works like Word, but it's free!

For the next 90 days or so, you can download a Chinese-made alternative to Microsoft's Office suite, for free. The download is only 15MB, and includes equivalents to Word, PowerPoint and Excel. The company offering the free software is Kingsoft, whose more poetic Chinese name is Gold Mountain (金山). Kingsoft's Office suite is named WPS Office.

The software is available in versions for both Windows and Linux. According to King Soft, this new release of WPS Office "is an overhaul of 5 million lines of code by 100 engineers over three years", with all the work done in China. The software is available for free for 100 days, after which it will be sold at prices designed not to compete with Microsoft Office, but with pirate versions of Microsoft Office.

There seem to be two main problems with Kingsoft's office suite:

- The menus are only available in Chinese, and there is no English spell check;
- The interface design is a clone of Microsoft Office's interface.

Problems aside, this is the the beginning of a trend that is going to make people in Seattle and Mountain View nervous. Just as the West Coast is getting used to the competition from India, cut-rate Chinese software, hand-coded in Beijing and Hangzhou, is going to present another challenge.