Archive for the ‘Search Engines’ Category

How google.com deal with China’s Great Firewall

Monday, July 2nd, 2007

GFW in China is some kind terrible “big brother” which judge the live or death of a web site. Beside some websites who have been executed permently(such as zh.wikipedia.com, etc.), GFW scan the content on the fly and once some content triggered the alarm, the current TCP connection will be reseted, and this site will be blocked from your current IP address for some minutes.

Google is one of sites sufered from GFW so much, google launched a self-censored site in China with the domain name google.cn, however, for most of the people, they already get used to “www.google.com” or they have toolbar installed with google.com, and non-technical people can’t tell from “google.com” to “google.cn”, still most of the traffic will go to Google.com, that is to say, under the monotoring of GFW.

Google developed some hack inside the front page of www.google.com, it’s a small pieces of Javascript, which send some ajax request and if such ajax request is not returned successfully, it will redirect the URL to google.cn with the same query strings. ( I guess only people from China IP will get this piece JS code, but not confirmed)

The JS code is not quite readable, obviously google don’t wish you do so…:) But we can still easy understand what it intend to:

<script>
<!–
window.google={kEI:”7dmIRuzeIJOesAKC0pnIDw”,kEXPI:”17259,17497,17585″,kHL:”en”};function sf(){document.f.q.focus();}
window.rwt=function(b,d,f,j,k,g,l){var a=window.encodeURIComponent?encodeURIComponent:escape,h=”",i=”",c=b.href.split(”#”),e=”";if(d){h=”&oi=”+a(d)}if(f){i=”&cad=”+a(f)}if(g){e=”&usg=”+g}b.href=”/url?sa=t”+h+i+”&ct=”+a(j)+”&cd=”+a(k)+”&url=”+a(c[0]).replace(/\+/g,”%2B”)+”&ei=7dmIRuzeIJOesAKC0pnIDw”+e+l+(c[1]?”#”+c[1]:”");b.onmousedown=”";return true};window.gbar={};(function(){function C(a,d,b){var c=”on”+d;if(a.addEventListener){a.addEventListener(d,b,false)}else if(a.attachEvent){a.attachEvent(c,b)}else{var g=a[c];a[c]=function(){var f=g.apply(this,arguments),h=b.apply(this,arguments);return f==undefined?h:(h==undefined?f:h&&f)}}};var k=window,u=k.location,A=u.search,z=u.protocol,j=document,q=”appendChild”,m=k.gbar,l=”",s,w=”",e,i,n,t=”wivnlmzbjpcoegz0tqfys”;function r(a){return escape(unescape(a.replace(/\+/g,” “))).replace(/\+/g,”%2B”)}function v(a){return i[a].firstChild.tagName==”A”}function o(a){return A.match(”[?&](”+a+”=)([^&#]*)”)}m.init=function(){var a=0,d,b=”affdom,channel,client,hl,hs,ie,lr,ned,oe,og,rls,rlz”.split(”,”),c=o(”as_q”),g=o(”q”),f=j.getElementById(”gbar”),h=k.navExtra;s=o(”near”);if(h){for(var y in h){var p=j.createElement(”div”);p[q](h[y]);p.className=”gb2″;f[q](p)}}i=f.getElementsByTagName(”div”);g&&(l=g[2])&&c&&(l+=”+”);c&&(l+=c[2]);while(v(a++)){}n=t.charAt(a-1);for(a=0;b[a];a++){d=o(b[a]);d&&(w+=”&”+d[1]+r(d[2]))}for(a=0;i[a];a++){v(a)&&B(a)}C(j,”click”,m.close)};function B(a){var d=i[a].firstChild,b=d.href+(d.href.match(”[?]“)?”&”:”?”),c=t.charAt(a);if(c!=”z”){b+=”tab=”+n+c;if(”com”.indexOf(c)>=0){b=b.replace(”http:”,z)}else{b+=w;if(l){b+=”&q=”+r(l);s&&n==”l”&&(b+=”+near%3A+”+r(s[2]))}}}d.href=b}function x(a,d,b){a.display=a.display==”block”?”none”:”block”;a.left=d+”px”;a.top=b+”px”}m.tg=function(a){var d=0,b,c,g,f=0;a=a?a:k.event;a.cancelBubble=true;if(!e){e=j.createElement(Array.every||k.createPopup?”iframe”:”div”);e.frameBorder=”0″;e.id=”gbi”;e.scrolling=”no”;e.src=”#”;j.body[q](e)}for(;n&&i[f];f++){c=i[f];g=c.className;if(g==”gb3″){b=c.offsetLeft;while(c=c.offsetParent){b+=c.offsetLeft}x(e.style,b,24)}else if(g==”gb2″){c.id==”gbar”+n&&(c.style.padding=”.2em .5em”);x(c.style,b+1,25+d);d+=20}}e.style.height=d+”px”};m.close=function(a){e&&e.style.display==”block”&&m.tg(a)};})();// –>
</script>

Does this work well?  Unfortunately not, and in most cases it does not work well.   Why, because the GFW ’s response is not as fast as expected.  In most cases, if the query result contain something GFW think “bad”, it will not able to stop you on first page, that is to say, the first result page will show correctly, but you will not able to go through the next pages.  Google’s Javascript make a call to check first, but that call may return correctly and then google get blocked, user will still get a error page and not able to access Google in the next several minutes.

Popularity: 15% [?]

Ask3D: only the best search engine frontend!

Monday, July 2nd, 2007

:) Pls. forgive this shameless self promotion.  I did have some small contribution on this coolest search engine frontend.  I can say in today’s search engine market, this frontend is still the most advanced.  Here is an Introducing Ask3D - A Truly New Way to Search.

You can get more details from Ask.com’s blog.

Our New Home Page: There’s No Place Like It

The power behind the panels

Yes, I should admit that the search result is not as good as it looks like, but I believe the backend guys are working hard to improve it. Image that, if such a cool frontend plus a similar strong bankend search engine, this should absolutly take some more pie in the search market.

Popularity: 16% [?]

Powerset, I just feel something don’t right…

Sunday, July 1st, 2007

Powerset, a legend “google killer”, everybody are expecting something fresh from them to make us suprise, including myself.

But I just felt something don’t right… I don’t know what it is and why I feel so, maybe just 6th sense? :) I write this blog as a record of what I think right now, and maybe in the future I will have chance to know what makes me think stupidand will be helpful to let me learn how to look forward something I may never imaged before.

Here is what I heard about Powerset (or rather say what I think they are after I read those buzz and subcribe to their Powerlab as a “Powerlabber”, maybe some mistakes):

1. They rasied a bunch of $$$ before the release anything to public;

2. Their website is as empty as google in the early years, except google can provide search they don’t;

3. They use Amazon’s EC2 and S3, so they don’t have to host a lot of servers and rent IDC racks;

4. They use Ruby on Rails to develop front end, they will have great guy to make Rails as fast as XXX; so they don’t worry about scaling;

5. Since I join Powerlab, I received nothing really interesting… The actual lab till now is still just a mail list, all I get was a private unaccessable Youetube video link, and finally when I can view it, it just tell me almost nothing except one well know screen shot.

So I just feel something is not right… but have no idea what’s not right and why I feel not right. Let’s just wait and see what Powerset guy will bring us and how to make us surprise.

Popularity: 18% [?]

Close
E-mail It