Google PageRank Checksum Algorithm

[doctorow]

Cooper, any idea how I can get the Yahoo beta toolbar? Would be interested to have a look at the WebRank feature...

[acidophil]

hi y-all!

yes, y!-WR whould be nice coz i think y! will be kick googles b...t soo.

but atm googles PR is relevant
and so i was wondering if someone has figured it out to get the checksum with a perl script???

btw has anyone re-checked the PRs, coz google tend to change them - delivering wrong PRs!

THX

acidophil

[Cooper]

Quoted From Google PageRank, Meet Yahoo WebRank
Where do the scores come from? Yahoo's help pages say the value is calculated by anonymously monitoring what people visit. So if many toolbar users visit particular sites, that might be a factor in helping the sites get a better Web Rank score.

But hold on! When I asked Yahoo earlier this month, I was told that the score is instead determined by Yahoo's search algorithm. That means visits to a site are not counted to create Web Rank. Instead, it's apparently link popularity based, as with Google's PageRank. Unfortunately, Yahoo hasn't yet corrected its help pages to reflect this.

A somewhat outdated article, but still pretty interesting. Jerry Rank ? LOL
However, so far, IMO Yahoo WR is practically useless . As I've heard in many places it does not show any rank for most of the toolbar (beta) users. So it's most likely that thier beta testing was primarily to keep the spies out. (I still wonder, how digitalpoint got thier hands on the webrank tool). The link you posted earlier was suppose to be the final know release of the beta version. However, it now downloads the normal toolbar. I tried it, but didn't quite convice me to keep it for more than 5 mins. Anyway, I had a few friends who were talking abt it back in april, so i guess they do have the beta versions. I'll check with 'em.

In the mean time, Google agents aren't uniform. I've been trying different stuff to accomplish high anonymity. However, unless run in client-side (like the firefox extension), it's hard to avoid sending unnoticeable requests from a single server (talking abt large volumes, like the PR tool at seo-guy). One option would be using socks. but thier availablity is totally undesireable. So I'm currently working with javascript to make the requests from the client side and obfuscating the html so that no one is aware of what's happening. (specially G)

A JSP/Servlet-Applet combination which has both server & client side control would be nice. But only a few hosts run TomCat. Btw, I'll try to write a perl version if I do get some time for it.

Also, for those who need the PR displayed on thier site, here's an easier way, http://www.pagerank.net/display-pagerank.php

P.S: I just heard that your script is circulating on P2P, some at a cost.

this is completely done with javascript requiring no backend php or cgi. it also seems to be platform independent. props go to here for the original javascript. it doesn't handle urls with parameters (like the url of this page in the thread), but i haven't seen any of the other implementations handle these correctly, either (they chop off the parameters). just prepend 'javascript' + ':'.

function%20zeroFill(a,b){var%20z=2147483648;if(z&a){a=(a>>1);a&=(~z);a|=0x40000000;a=(a>>(b-1));}else{a=(a>>b);}return%20a;}function%20mix(a,b,c){a-=b;a-=c;a^=(zeroFill(c,13));b-=c;b-=a;b^=(a<<8);c-=a;c-=b;c^=(zeroFill(b,13));a-=b;a-=c;a^=(zeroFill(c,12));b-=c;b-=a;b^=(a<<16);c-=a;c-=b;c^=(zeroFill(b,5));a-=b;a-=c;a^=(zeroFill(c,3));b-=c;b-=a;b^=(a<<10);c-=a;c-=b;c^=(zeroFill(b,15));return%20Array(a,b,c);}function%20GoogleCH(url){var%20length=url.length;a=b=0x9E3779B9;c=0xE6359A60;k=0;len=length;while(len>=12){a+=(url[k+0]+(url[k+1]<<8)+(url[k+2]<<16)+(url[k+3]<<24));b+=(url[k+4]+(url[k+5]<<8)+(url[k+6]<<16)+(url[k+7]<<24));c+=(url[k+8]+(url[k+9]<<8)+(url[k+10]<<16)+(url[k+11]<<24));mixvar=mix(a,b,c);a=mixvar[0];b=mixvar[1];c=mixvar[2];k+=12;len-=12;}c+=length;switch(len){case%2011:c+=(url[k+10]<<24);case%2010:c+=(url[k+9]<<16);case%209:c+=(url[k+8]<<8);case%208:b+=(url[k+7]<<24);case%207:b+=(url[k+6]<<16);case%206:b+=(url[k+5]<<8);case%205:b+=(url[k+4]);case%204:a+=(url[k+3]<<24);case%203:a+=(url[k+2]<<16);case%202:a+=(url[k+1]<<8);case%201:a+=(url[k+0]);}mixvar=mix(a,b,c);return%20mixvar[2];}function%20strord(string){var%20result=new%20Array();for(i=0;i<string.length;i++){result[i]=string.charCodeAt(i);}return%20result;}var%20ch=GoogleCH(strord('info:'+location.href));location.href='http://google.com/search?client=navclient-auto&ch=6'+ch+'&features=Rank&q=info:'+location.href

Originally Posted by rayg

this is completely done with javascript...

somehow a few spaces were automatically inserted in the code- they need to be stripped.

I was looking for this thing from some time thanks a lot all of u people.


Deepesh Agarwal,
http://www.freeware-alternative.uni.cc/
Visit website for Award-Wining Freeware's like: -
Easy Unistaller, The Ad-Police (Updated) - Now Blocks 14000+ Ad-Servers ,Service Controller XP(Updated), Dr.System 2.1 (New Build), Easy CDTray , Quick Deployment Wizard 1.0 - A Easy to Use Free Setup creator with High Compression, and Many More.....

[acidophil]

Originally Posted by rayg

somehow a few spaces were automatically inserted in the code- they need to be stripped.

*gmpf*
i tried to skip the %20 and rebuild some LFs but i couldn't get it working...

ok i don't imported it as a sniplet to mozilla mac (on my old littl' black (pismo)) but i thought i could use it as an inline js with document.write

is the source somewere D/L-able? if possible as .js too

THX
acidophil

pS.: searching for a perl script computing ch and an actionscript too!

this code really goes, but there is one mistake.

the right headers looks like this toolbarqueries.google.com not www.google.com its difference!

reply to xxxvendor@email.cz

Somebody on http://blog.outer-court.com/forum/ asked for a VBScript version. I posted my take at it here:

http://tumanov.com/projects/scriptle...nkchecksum.asp

[barranquilla]

I tried to port the checksum algorithm to Java and I had some partial success. However, I think there's a problem with the way Java manages hex numbers (unsigned longs, maybe?). Here's the source:

import java.io.*;

public class GoogleRank {

public int GOOGLE_MAGIC = 0xE6359A60;

// Google Rank
private long zeroFill(long a, long b)
{
long z = 0x80000000;

if ((z & a) == z)
{

a = (a>>1);

a = a & (~z);

a = a | 0x40000000;
a = (a >> (b-1));
}
else
{
a = (a >> b);

}

return a;
}

public long[] mix(long a, long b, long c) {
a -= b; a -= c; a ^= (zeroFill(c,13));
b -= c; b -= a; b ^= (a<<8);
c -= a; c -= b; c ^= (zeroFill(b,13));
a -= b; a -= c; a ^= (zeroFill(c,12));
b -= c; b -= a; b ^= (a<<16);
c -= a; c -= b; c ^= (zeroFill(b,5));
a -= b; a -= c; a ^= (zeroFill(c,3));
b -= c; b -= a; b ^= (a<<10);
c -= a; c -= b; c ^= (zeroFill(b,15));
long[] mix = new long[3];
mix[0] = a;
mix[1] = b;
mix[2] = c;

return mix;
}

public long GoogleCH(String url) {

long a = 0x9E3779B9;
long b = 0x9E3779B9;
long c = GOOGLE_MAGIC;
int length = url.length();
int len = length;
int k = 0;


while(len >= 12) {
a += (url.charAt(k+0) +(url.charAt(k+1)<<8) +(url.charAt(k+2)<<16) +(url.charAt(k+3)<<24));

b += (url.charAt(k+4) +(url.charAt(k+5)<<8) +(url.charAt(k+6)<<16) +(url.charAt(k+7)<<24));
c += (url.charAt(k+8) +(url.charAt(k+9)<<8) +(url.charAt(k+10)<<16)+(url.charAt(k+11)<<24));

long [] mix = mix(a,b,c);
a = mix[0]; b = mix[1]; c = mix[2];
k += 12;
len -= 12;
}

c += length;
switch(len) /* all the case statements fall through */
{
case 11: c+=(url.charAt(k+10)<<24);
case 10: c+=(url.charAt(k+9)<<16);
case 9 : c+=(url.charAt(k+8)<<8);
/* the first byte of c is reserved for the length */
case 8 : b+=(url.charAt(k+7)<<24);
case 7 : b+=(url.charAt(k+6)<<16);
case 6 : b+=(url.charAt(k+5)<<8);
case 5 : b+=(url.charAt(k+4));
case 4 : a+=(url.charAt(k+3)<<24);
case 3 : a+=(url.charAt(k+2)<<16);
case 2 : a+=(url.charAt(k+1)<<8);
case 1 : a+=(url.charAt(k+0));
/* case 0: nothing left to add */
}
long[] mix = mix(a,b,c);
/*-------------------------------------------- report the result */
return mix[2];


}

public static void main(String[] args) {

try {
GoogleRank g = new GoogleRank();
long checksum = g.GoogleCH("info:http://www.example.com/");

// Anything go wrong?
} catch (Exception f) {
System.out.println("Exception: " + f.toString( ));
}
}
}


Any pointers on what may be wrong are very appreciated...

Originally Posted by Unregistered

Windows tool based on above's code:

http://www.my-eldorado.net/GGChecksumRecovery.zip

This has been moved to...
http://www.my-eldorado.net/GGChecksumRecovery_win32.zip

The also posted the great PHP script you guys put together, though sadly declined to include you, the authors names...
/*
This code is released unto the public domain
*/

While I'm posting, I just wanted to say:
1) Great to see 3 of you working together to produce all this code.
2) That I respect and admire you all, for this great work.
3) Reading your collaborative efforts has made me decide to learn programming. I realise I want to "play".

You guys all rock and it has been great reading this long thread.
Well Done.

- David (Australia)

My PageRank Calculator gets back.. but with a new name and new features.
Now it's GoogleGetInfo and gets PageRank and 10 parameters more.

Check it at http://dhost.info/zeni/google-info

It gets all the parameters from Google in XML and returns them in an array with those parameters:

//0: URL
//1: Title
//2: PageRank. The Most Wanted!
//3: Last Crawled Date. Syntax: " 31 Aug 2004". Note the first space, use trim() to remove it.
//4: Page Text that appears in Google
//5: Language of the page (2 letters)
//6: URL in the Directory. Get to http://www.google.com/<this> to see it.
//7: Title in the Directory.
//8: Description in the Directory.
//9: File Size. Ex: "59k"
//10: Looks like a checksum of the page, not sure

Read googlegetinfo.php comments for info.

Note: The PageRank calculator URL (http://zenitram.th4y.com/pagerank) won't work anymore, the server disappeared...

this php's program won't certainly work for a long time : google introduced yesterday a new version of the toolbar (v 2.0.114) and changed the checksum algorithm
Maybe a little change, maybe not...
Can someone try to decompilate and see ?

Mosert

for information, the 4 last version of the toolbar were :
* version 2.0.111*: 2004-05-14
* version 2.0.112*: 2004-07-15
* version 2.0.113*: 2004-07-30
* version 2.0.114*: 2004-09-09

[doctorow]

Mosert, can you show me how the checksum has changed in 2.0.114 since 2.0.113- like show me the CH of the same URL for both versions?