Expand
CollapseTue 08 January, 2008
[[ This is a content summary only. Visit my website for full links, other content, and more! ]]
Mon 07 January, 2008
Here’s something I missed while I was offline last week. It looks like Matt Marlon, who served as CEO of Traffic Power, has landed himself in jail. I have mixed feelings about this, but the mixed feelings are mostly satisfaction, curiosity, and just a smidge of glee. 
In my opinion, the stuff that TrafficPower was doing was just bad news.
I think the search engine optimization industry has made a lot of progress in the last few years. It’s a little less common to get cold calls from SEOs that guarantee #1 rankings but won’t tell you how they try to do it. And if a large SEO company wants to try something high-risk with a client, they’re more likely to explain the potential risks to that client first. There are still issues, of course, but I was looking over a list of 20+ blackhat SEO companies that I compiled back in 2002. The majority either went out of business or have transformed into white-hat SEO companies.
For a while now, I’ve had a slight hunch that clients that embrace blackhat SEO on their site are willing to cut corners in other areas of business as well. Earlier today I was reviewing an email from 2001 (!) where Google removed a very large company’s website from our index for hidden GIF links, machine-generated doorway pages, and cloaking. It’s interesting to look back with the benefit of hindsight now. Later on, the company:
- had 10+ employees convicted for inflating revenue
- the CEO was sentenced to 10+ years in jail
- another executive was sentenced to 2+ years in jail
Can I definitively claim that there’s a connection between a willingness to embrace blackhat SEO and a willingness to cut corners in other areas of business? No, of course not. But I have seen several examples like the one I mention above. That’s why I’m glad that as more site owners learn about SEO, the long-term odds of blatent SEO scams going undetected go down.
Hat tip to Aaron Wall for mentioning Matt Marlon’s new location (jail) to me. Aaron and I have had differences, but I give him major points for fighting Traffic Power when they sued him. That led to one of the first times that Google confirmed a webspam penalty in public in order to dispute one of the lawsuit’s claims:
I can confirm that Google has removed traffic-power.com and domains promoted by Traffic Power from our index because of search engine optimization techniques that violated our webmaster guidelines at http://www.google.com/webmasters/guidelines.html. If you are a client or former client of Traffic Power and your site is not in Google, please see my previous advice on requesting reinclusion into Google’s index to learn what steps to take if you would like to be reincluded in Google’s index.
Ah, good times.
原文：Improve snippets with a meta description makeover
发表于：2007 年 9 月 27 日，星期四，下午 6 时 46 分
页面摘要是我们为每个搜索结果显示的简短的文字预览。你的页面摘要的质量直接影响到你的网站被点击（即谷歌带给你的流量）的可能性。我们使用了多种策略选择页面摘要。对你来说，如果你能为每个 URL 写一段信息丰富的元描述，你就可以控制谷歌对你的 URL 的页面摘要的显示。
<META NAME="Description" CONTENT="把一段信息丰富的描述放在这里">
为什么谷歌关注元描述?
我们希望页面摘要能准确地代表搜索结果。我们总是愿意尽可能地显示页面的元描述，因为它为用户提供该网址内容的清晰概念。这会指引用户更快地找到好的结果，减少点击后走回头路而使他们感到厌烦并无端增加网络流量的可能性。请记住，由关键词堆砌而组成的元描述不能达到这个目标，且不太可能被显示在正常的，非元描述的页面摘要。另一个值得注意的是，准确的元描述可以增加点入机会，他们不会影响你的页面在搜索结果中的排名。
高质量的元描述页面摘要
低质量的元描述页面摘要
什么是好的元描述策略？
对不同页面给予不同的描述
不要对每一页都采用相同或相似的描述，这样对单个页面出现在网页搜索结果的效果没有帮助。如果你这样做，我们不太可能显示你的样板文字。你应为每个具体的一页写一个准确地描述。在首页或其他总结性页面上给一个网站级别的描述，而对其他单个页面使用页面级别的描述。如果你没有时间为每一个单页去创造一个描述，显然你应该对你的不同网页考虑优先级。至少，为你的关键 URL（譬如你的首页和热点页面）创建一个描述。
在描述中明确包含标记性事实
元描述不一定是一个句子格式。它是一个放置关于当前页面结构化数据的好地方。举例来说，新闻或博客帖子可以列出作者，出版日期，或署名信息。这给潜在的访客以非常相关的信息。如果不放在元描述里，它们就可能不会显示在页面摘要里。同样，在产品介绍的页面可能有关键的信息（价格，生产日期，制造商等)，他们可能分布在当前页的各处，所以不可能通过一个页面摘要来显示所有这方面的资料。如果你把它们放在元描述里，页面摘要就能显示它们了。比如，看一看下面的取自一个主要产品聚集商为“哈利波特7”一书的元描述。
不是很理想的描述：
<META NAME="Description" CONTENT="[域名：redacted]: 哈利波特(7)：死神的圣物: 书籍: J. K. Rowling,Mary GrandPré 作者： J. K. Rowling,Mary GrandPré">
有很多原因将导致这段元描述对我们的搜索结果来说不是一个好的页面摘要：
* 描述中的书名信息和该页面的标题有完全重复的信息。
* 信息描述本身就有重复（J. K. Rowling, Mary GrandPré 被罗列了两次）。 .
* 描述提供的信息不清楚，谁是 Mary GrandPré?
* 名字间少了个空格，用了太多的冒号使得描述很难读。
一般用户看谷歌的搜寻结果页可能只花一秒钟来扫描某一页面摘要，上述的所有因素都意味着用户很可能跳过这一结果。考虑以下的元描述：
好多了的描述:
<meta name="Description" content="作者: J. K. Rowling, 插图: Mary GrandPré, 分类: Books, 价格: $17.99, 长度: 784 页">
什么改变了呢？没有重复，更多的信息，所有信息都清楚地标记和分开。并不需要真正的额外工作来产生这样的一种质量：价格和长度是仅有的新数据，他们已经在当前页显示了。
用程序来产生描述
对诸如新闻媒体源之类的网站来说，为每一页产生一个精确而独特的描述很简单：因为每一篇文章都是手写的，它需要极少的努力来添加了一个句子描述。对诸如产品聚合器之类的较大型数据库驱动的网站来说，人工写描述就比较困难了。在这种情况下，使用程序来产生描述是适宜的并受到鼓励的。只是你要确保你的描述看上去不是在产生网络垃圾。就像我们在上面第一点谈到的，好描述的可读性强并且是多样的。在我们上面谈到的第二点中的网页特有的数据也是一个使用程序产生描述的很好的候选方案。
使用高质量的描述
最后，请确保你的描述确实是在描述你的页面。人们很容易对元描述的质量控制产生松懈情绪，因为他们在为访客的用户界面上不是直接可见的。但是元描述有可能显示在谷歌搜索结果中，如果描述的质量足够高。在元描述上的一点点额外工作可以大大地提高搜索结果中的页面摘要的相关性。这不但可能改善你的网站的质量，也有可能增加你的用户流量。
Crontab 主要是在做排程, 通常一般寫法大概都是如下:
0 0 * * * /usr/local/www/awstats/cgi-bin/awstats.sh
這種寫法規則如下:
分 時 日 月 年
對應表如下:
field allowed values
----- --------------
minute 0-59
hour 0-23
day of month 1-31
month 1-12 (or names, see below)
day of week 0-7 (0 or 7 is Sun, or use names)
看到彥明長輩寫才知道有這種寫法:
@hourly /usr/local/www/awstats/cgi-bin/awstats.sh
使用 @hourly 對應的是 0 * * * *, 還有下述可以使用:
string meaning
------ -------
@reboot Run once, at startup.
@yearly Run once a year, "0 0 1 1 *".
@annually (same as @yearly)
@monthly Run once a month, "0 0 1 * *".
@weekly Run once a week, "0 0 * * 0".
@daily Run once a day, "0 0 * * *".
@midnight (same as @daily)
@hourly Run once an hour, "0 * * * *".
特別是看到 @reboot, 所以寫 rc.local 的東西, 也可以使用 @reboot 寫在 crontab 裡面?... XD
找時間再實驗看看. 更多詳細的應用, 可以 man 5 crontab :)
Yahoo! co-founder and CEO Jerry Yang was on stage today at CES and he showed a fascinating glimpse into the future of Yahoo. Yang outlined a product strategy that takes the simplicity and all-in-one portal approach that Yahoo! is famous for - and pushes it into the digital life arena by utilizing email, social networking, mobile and widgets. In Yang's words, Yahoo! aims to be the "most essential starting point for your life". The name for this new product suite will be Life!.
Dan Farber has a great write-up of Yang's speech on ZDNet, so in this post I'll summarize that and give some analysis on Yahoo's future.
The crux of the new Life! product is to unite Yahoo Mail, MyYahoo, search and the Yahoo home page in a social networking context. This will be tied together using technologies like structured data, tagging and open standards like OpenID (see Marshall's post this morning on Flickr adopting OpenID).
Dan Farber reported:
"Yahoo Mail (communication services) serves as the hub, but the interface also includes third-party applications and social context. Connections are contacts, and based on frequency and volume of communications email is reordered on the strength of the connections. The page also includes updates from connections, showing what is relevant rather than just relatively static inbox."
Interesting to note here that Google is also looking to use Web email as a hub for social networking. So both Yahoo! and Google are really targeting email to get users attention and integrate social web products.
Image: Dan Farber, ZDNet
Zimbra Integration
Yahoo! will integrate Zimbra into Life!. Zimbra is an open source mail and messaging platform, which became famous in the web 2.0 world for its mashups ability. It was acquired last year by Yahoo; and so it's exciting to know that Zimbra will be utilized more fully in Yahoo's product range.
Yahoo Go 3.0 - Widgets Galore
Finally, Yang showed off the new version of its Yahoo Go mobile platform. Labeled Yahoo Go 3.0, it will introduce widgets for things like email and Flickr, enabling those Y! products to be better utilized on mobile phones. Go will also become an open platform for third party widget developers - see our coverage of this last night. As Marshall wrote, unlike Google's Android OS, the Yahoo! Go platform will work on more than 250 mobile devices that Go already works on.
Yahoo has partnered with the likes of eBay, Viacom and News Corp on widgets for Go. Yahoo is also working with LG Electronics and other phone vendors to make widgets run natively on phones.
From the user side, a new widgets gallery called MySnippets will give users feeds and services that can be embedded in Yahoo pages. A beta version of 3.0 will be available later today by directing your phone’s browser to http://beta.m.yahoo.com
Conclusion
It looks like Yahoo is finally beginning to tie together its many 'best of breed' web 2.0 products (Flickr, Zimbra, Mail, Go) into an integrated product offering. Mobile has been an area of development at Yahoo for some time - they appear to be doing more work than Google and Microsoft in the mobile portal / social networking arena, and with Go 3.0 it might crack the mainstream in a big way. Yahoo has also been a leader in terms of open standards - see our coverage of Ian Rogers about the Media Web for one example.
Time will tell whether these moves by Yahoo will get them out of the funk they fell into during 2007, but Yang's speech at CES had many encouraging signs.
Shanti Braford details how his Ruby on Rails based website survived a 24 hour 550,000+ pageview digg attack. His post cleanly lays out all the juicy setup details, so there's not much I can add.
Hosting costs $370 a month for 1 web server, 1 database server, and sufficient bandwidth. The site is built on RoR, nginx, MySQL, and 7 mongrel servers. He thinks Rails 2.0 has improved performance and credits database avoidance and fragment caching for much of the performance boost.
Keep in mind his system is relatively static, but it's a very interesting and useful experience report.
Flight Thru Instruments is a US Navy pilot-training manual from the 40s designed by the Graphic Engineering Staff at General Motors. The information graphics are stunning, colorful, and clear.
Here’s the full Flickr set. Hat tip: Airbag.
Flight Thru Instruments is a US Navy pilot-training manual from the 40s designed by the Graphic Engineering Staff at General Motors. The information graphics are stunning, colorful, and clear.
Here’s the full Flickr set. Hat tip: Airbag.
一个叫张◇◇小学生，13岁，在接受中央电视台采访的时候，对着镜头说了一句现在的网站“很黄，很暴力”，于是这几天就成了新闻。开始我还不知道，后来发现有人在我博客上留言，好几个人留言说“很黄很暴力”，我的直觉告诉我，这肯定有是一句流行语，我统统删掉，如果自己不会说话尽量不要说别人嘴里吐出来的东西。
后来一个同学给我发邮件，详细告诉我“很黄很暴力”的渊源，我这才明白是怎么回事。
曾经有人说我写博客很黄很暴力，比如常说脏话，而且不回避使用男女生殖器官名词，还有经常用一些含有性意向的词句；还有说我写东西不讲道理，用句成语就是强词夺理，骂人非常刻薄，充满了语言暴力。还好，张◇◇小朋友不是看了我的博客之后发出这样的感慨的。看了我博客发出类似张◇◇感慨的都是好人，正人君子，听妈妈的话和党的话长大的好孩子，所以我也能理解。如果我的文字（包括图片）引起了您感官上的不适，那我在这里说声对不起，建议您以后不要再来，万一诱发您的违法犯罪，万一再成为CCTV-1焦点访谈的素材，多不好啊。但我这种写字的方式不会改变，我会一如既往，我想说什么就说什么，看清楚，不是您想让我说什么我就说什么。如果您非要我说什么，我以前也表过态：请付费吧。
言归正传，说说张◇◇的事情。事情既然从CCTV引发的，那我们先从CCTV说起。这个新闻我没看，其实根本不用看，看了这么多年CCTV，还不知道他们的路数——就是编导或制片人希望听到一句他们想听的话，用诱供的方式逗出被采访者说出这句话。如果仔细分析，每句话都可能有代表性，但要看你怎么说，看你把什么当成重点。你看做电视的人都这样，采访一个人可能录像一个小时，但最后用的时候可能就几句话，选择这几句话很讲究的，被采访者必须说出编导和制片人的心声，比如张◇◇小朋友就做到了，然后掐头去尾，断章取义。其实央视的记者完全可以上各大门户网站，把页面打开，对着电脑屏幕拍一些很黄很暴力的页面，但是他们为了更有震撼力，一定要把一个未成年人揪出来对着镜头说这句话，而且说得那么精辟、深刻。这五个字充分体出了中央电视台的伪善新闻价值观，同时也暴露了所有中国人都习以为常的传统——意识不到保护未成年人。
让一个未成年人站出来说一些与她关系不大的话，并且还要承载着某种道义上的东西，CCTV真鸡巴缺德。是的，如果让我或者宋祖德老师说这样的话，大家一定以为是在恶搞，让一个单纯且未成年的小姑娘说这样的话，就很有杀伤力，足以让人们反思互联网，这是央视编导和制片人的想法之一，想法之二是这样做可能获得更多的道德分数。
事实上，我没有看到反思，我看到了暴力，我没有看到央视获得什么道德分数，却看到了他们的无知和不负责任，虽然中国的《未成年人保护法》里面没有规定未成年人不能接受媒体采访，但是从未成年人保护的角度出发，也该慎之又慎，甚至我认为，CCTV把张◇◇的名字打出来都算侵犯她的隐私。土摩托为他第一次上电视感到骄傲和自豪，他也是在13岁左右的时候上电视，他自豪的是参加了“智力竞赛”，20年后，张◇◇作为一个成年人在回顾她第一次上CCTV的情景，是否会觉得是个噩梦呢，会有土摩托这样的自豪感吗？
网民们像是被捅到了G点，立刻兴奋了起来，当他们拿这件事去说事，去恶搞，也都忽略了一个对未成年人保护的意识（他们又能有什么意识呢）。可见，对未成年人保护在中国是长路漫漫。而众多网民狂欢式的恶搞，实际上在伤害着一个未成年人的心灵，她的图片遍地都是。他们图的是过瘾，比如很兴致勃勃地宣称这五个字是今年第一条流行语，妈了逼的，当然这个小朋友不是你女儿或你侄女。
中央电视台、众多网民、张◇◇的父母都忽视了对未成年人保护的意识。我们的青少年的成长已经很不健康了，看看87.65%“独一代”心灵的扭曲（你会问这数据怎么这么准啊？我他妈能给你们“独一代”反驳我的借口吗？你不同意就把自己列入13.35%里面），再看看教育的毒害，在互联网时代，知识越多越弱智。对未成年人再不想点办法保护，大概就完蛋了。我想，类似张◇◇这样的孩子在成年人无意识下受到伤害的事情有很多，人们真的习以为常了，所以才有“CCTV很黄很暴力”的事情发生。
曾经有个朋友让我把她的一个亲戚（10岁左右的小孩）的博客链接加在我的博客上，被我拒绝，首先，我觉得让未成年人开博客本身就是家长该慎重考虑的事情，是否会影响孩子的成长？其次，我绝对不能容忍在我的博客里为未成年人的博客做链接，因为我担心会有很多傻逼去伤害未成年人；再次，中国的互联网和中国电影一样都面临这样的一个问题：内容该分级。比如新浪网站，我一直是把它当成色情网站来看，就是把男盗女娼和仁义道德完美结合在一起的那种门户网站。当然，其他网站也好不到哪里去，有一个算一个。这就好比把妓院开在超市里面一样，谁都能看到。如果网络不分级，好多问题都解决不了。有了分级制，至少可以加强公众对未成年人保护的意识。成年人，怎么都好说，而未成年人，怎么都不好说。
I got Designing Interactions over the holidays and have really enjoyed it so far. I’m not done with it, but I’m done enough to be able to recommend it.
The book looks at the evolution and breakthroughs of interface and industrial design over the last 40 odd years.
The history traces a squiggly line through the invention of the mouse, the graphical user interface, the PC, the laptop, the handheld, gaming, online services, the internet, and beyond. It’s a really interesting look at how this builds on that and that changes this.
The 700-page book is richly illustrated and includes interviews with some of the world’s most important figures in design, product development, design education, and computer science.
(Thanks for the book, JD)
I got Designing Interactions over the holidays and have really enjoyed it so far. I’m not done with it, but I’m done enough to be able to recommend it.
The book looks at the evolution and breakthroughs of interface and industrial design over the last 40 odd years.
The history traces a squiggly line through the invention of the mouse, the graphical user interface, the PC, the laptop, the handheld, gaming, online services, the internet, and beyond. It’s a really interesting look at how this builds on that and that changes this.
The 700-page book is richly illustrated and includes interviews with some of the world’s most important figures in design, product development, design education, and computer science.
(Thanks for the book, JD)
New York Magazine
New York Magazine’s Agenda calendar shifts focus depending on which day you mouseover.
WellStyled.com
Check out how the two languages are swappable at WellStyled.com. The sidebar is in Czech but if you click it, it swaps with the English. Even the nav changes too. And it sticks as you move through the different pages.
Bento
Here’s how Bento combines Address Book show and edit states.
Have an interesting link, story, or screenshot for Signal vs. Noise? Contact svn [at] 37signals [dot] com.
New York Magazine
New York Magazine’s Agenda calendar shifts focus depending on which day you mouseover.
WellStyled.com
Check out how the two languages are swappable at WellStyled.com. The sidebar is in Czech but if you click it, it swaps with the English. Even the nav changes too. And it sticks as you move through the different pages.
Bento
Here’s how Bento combines Address Book show and edit states.
Have an interesting link, story, or screenshot for Signal vs. Noise? Contact svn [at] 37signals [dot] com.
Wikia允许搜索者添加或修改跟关键词相关的简短说明,叫做mini-article简述定义,可以包含文本和链接,其他人搜索相同关键词时这条编辑将出现在搜索结果最前面,现在貌似没有针对作弊有效措施,只有警告,封禁账号/网址的惩罚,呵呵.
搜索结果中会显示url排名的权值信息和页面分析算法,采用nutch/Lucene提供服务.正如官方所述,目前的搜索质量还不高。而Wikia四大原则透明\社群\品质\隐私,似乎目前看来都不是搜索引擎的核心使用价值哦.目前看来就凭着这个叫板Google和Yahoo还需要时日.
EXECUTE DBMS_STATS.CREATE_STAT_TABLE ('SCOTT', 'STATTAB','SYSAUX');
在 SYSAUX 表空间上创建 STATTAB 用以存储统计信息, 所有者是 SCOTT 用户。
导出统计信息. (在任何可能更改表的统计信息的 DDL 操作之前, 一定要导出统计信息)
EXEC dbms_stats.EXPORT_SCHEMA_STATS (ownname=>'scott',stattab=>'stattab',STATID=>'foo_20080107');
这里建议手动设定一下 STATID. STATID 命名规则建议用 对象名(SCHEMA名)+ 时间(注意粒度).
至于导入整个 SCHEMA 的信息，一定要慎重再慎重。
在任何可能更改表的统计信息的 DDL 操作之前, 导出(备份)统计信息
EXEC dbms_stats.export_table_stats (OWNNAME=>'scott',TABNAME=>'foo',STATTAB=>'stattab',STATID=>'foo_20080107');
恢复该表的统计信息(之前要导出当前的统计信息):
EXEC dbms_stats.import_table_stats (OWNNAME=>'scott',TABNAME=>'foo',STATTAB=>'stattab',STATID=>'foo_20080107');
为了避免误导，需要说明的是，我只收集表和索引的统计信息。尽量不用 DBMS_STATS 收集统计信息，要问为什么? 去看看 DBMS_STATS 相关的 Bug 就知道了(比如飞龙说的这个问题)。只有在 ANALYZE 力有未逮之时才会考虑用 DBMS_STATS.
这里说的 和 ADDM 无关，建议在熟知 ADDM 之前，最好别用这玩意儿。
--EOF--
这是我的 Oracle 11g 系列的文章之一(这一篇没啥含量，算是笔记吧)。
在 11g 之前的版本中，DBMS_STATS 自动统计收集(Automatic Statistics Gathering)默认的阀值是 10%, 这个 10% 是不可以修改的。这对千变万化的企业数据库来说环境来说，有些死板，如果是个超大的表，默认的 10% 数据也是海量了，会把整个资源占死。Oracle 11g 中，这个属性可以通过修改 STALE_PERCENT 属性来修改, 有全局（DBMS_STATS.SET_GLOBAL_PREFS ）和表级别（DBMS_STATS.SET_TABLE_PREFS）两种。
例子语句:设定:SQL> exec dbms_stats.set_table_prefs(null,'EMP','STALE_PERCENT',1);
修改为 1％. 范围从 1-100.
恢复:
SQL> exec dbms_stats.set_table_prefs(null,'EMP','STALE_PERCENT',null);
查询:
SQL> select dbms_stats.get_prefs('STALE_PERCENT',null,'EMP') from dual;
虽然我很少用 DBMS_STATS 收集数据对象统计信息，不过它的有些用途还真的不错。
Refer: Metalink Note:390737.1 --EOF--
Sun 06 January, 2008