目前只对经常上国外网站的用户有效。。。。很简单。。。点击下面的按钮就行了:(如果是RSS阅读器里看到这篇blog请点击这个链接在新窗口打开本篇博文)
如果你在国外网站、特别是新兴Web 2.0网站混得相当多的话,这个结果还是非常准确的。转载至Mike On Ads的博文Using your browser URL history to estimate gender。大概原理是:在你访问过的常见网站中,有一个男女比率,例如thepiratebay.org是个非常男性的网站,男女比是2.13;而facebook的男女比是0.83。然后用Quantcast统计的top 10k网站男女比例,算出你是男是女的可能性。也就是利用浏览器的历史记录来分析你的性别。你可能会问:“Javascript能直接访问我浏览器的历史记录么?”
其实是SocialHistory.js在起作用。这是一个可以得知是否访问过一个网站的Javascript库。原理是什么呢?一般来说,我们的浏览器或者CSS会以颜色区分未访问的链接和访问过的链接。那么我们可以iframe一个页面,包含一个链接例如 www.dormforce.net。然后用js获得这个链接的颜色,就知道你是否访问过www.dormforce.net这个网站了。(强大吧。。。而且跨所有浏览器,包括IE、FF)。这个原理好囧啊。跟用mmmmmmmmmmmmmmmmmmm这一串字符的像素长度判断你电脑上是否安装某一个字体的那个js库 JavaScript/CSS Font Detector 一样。都是很囧的js技术
对您的性别是否判断准确呢?欢迎在下面评论,或者到原作者的blog、reddit、digg去看看其他人是怎么看的。
摘录囧评几条:
- WTF Im 76% likely to be a female? Good thing I'm on my mom's computer..
- i find its more accurate to spread my visitors' legs and look at their genitals. it works most of the time. 大意就是这js还不如直接看JJ来的快。。。
- I was surprised to see that the most female site that I'd visited was barackobama.com. 0.68 : 1, was not expecting that.
- Trust me, if you clicked on that link, you are a guy. no algorithm needed. (如果你被javascript关键字骗进来,并且点击了那个按钮,那么就不用测了。你100%是个男的。。。。)
其实呢。。我们也可以在搞一个天朝版的啊。。。Qzone男女比:0.65;xiaonei:0.7;J5团:0.8;tianya:1.1;Ubuntu.org.cn:2。。。哈哈。。数据是我乱写的。
标签: css, javascript, sns
54%:46%
我觉得还不如给我一个1:1......
Likelihood of you being FEMALE is 50%
Likelihood of you being MALE is 50%
WTF
92% male…
哈,准的
Likelihood of you being FEMALE is 26%
Likelihood of you being MALE is 74%
Site Male-Female Ratio
msn.com 0.92
sourceforge.net 1.74
rarlab.com 1.74
Likelihood of you being FEMALE is 21%
Likelihood of you being MALE is 79%
挺准:)
Likelihood of you being FEMALE is 65%
Likelihood of you being MALE is 35%
WTF...
Likelihood of you being FEMALE is 50%
Likelihood of you being MALE is 50%
居然是不男不女的。