Google PageRank Checksum Algorithm

correction; that should be %2F, not %2

%2F is a slash, see http://www.asciitable.com/

Without a trailing slash it isn't a proper URI

Alex & co's calculator:
url: http://www.example.com/
ch: 62863130722

Homelinux's calculator
url: http://www.example.com/
ch: 6540747202

Spot the huge difference?

I get different values for all domains I've tryed. Why? Wasn't your goal to create a free version of theirs?

[alexstapleton]

http://alex.vort-x.net/google/
It works.

url: http://www.example.com/
ch: 6540747202

See. Any problems your having are due to your server setup. Not the code. Post information on that and we might be able to help. PHP version, whether you modified the script or not, CPU, web server (if any.)

[alexstapleton]

Originally Posted by Unregistered

I just developed a complete all-in-one function in PHP that using The Google Checksum Calculator, by Alex Stapleton, Andy Doctorow, Vijay "Cyberax" Bhatter, and a few others, (https://www.mobileread.com/forums/sho...9&postcount=87) and XMLize.php by Hans Anderson (http://www.hansanderson.com/php/xml/) automatically calculates the PageRank for a page. Try it at:
http://zenitram.th4y.com/pagerank
And the source code is at:
http://zenitram.th4y.com/pagerank/source.txt
Hope you like it

iirc it breaks the google TOS and they can block you for using it?

One nice thing about this is that people won't be able to sell this code for $$$ anymore

Originally Posted by alexstapleton

http://alex.vort-x.net/google/
It works.

url: http://www.example.com/
ch: 6540747202

See. Any problems your having are due to your server setup. Not the code. Post information on that and we might be able to help. PHP version, whether you modified the script or not, CPU, web server (if any.)

Excuse me if I don't see your point, but I tested yours and the payable versions online so that's why I wondered. Because if you both get different ch's then Only one of the values should be able to use when retreiving the PR

i don't think they'll block me, there are 1000000s of requests for the Google Toolbar every day, and they can't determine what ones are from my script and what ones not.

However they can change the algorithm, but they need to update all the Google Toolbars out there before they can remove access via the actual algorithm

[alexstapleton]

they can block requests coming from zenithram.th4y.com. so it is perfectly possile, and easy, for them to block anyone that breaks their TOS.

they can change the algotithm, and all the toolbars automatically, transparently update. but what benefit would that give them?

[alexstapleton]

Festiz: If you get different results, then your doing something wrong. Because it works and produces the same values as the one at the homelinux address. There is a pagerank fetching form on my website now so you can test that the values work.

Updated by Alex and Bill Zeller.

PHP Code:

<?php
header("Content-Type: text/plain; charset=utf-8");
/*
* Written by Alex Stapleton
*
* With thanks to Vijay "Cyberax" Bhatter
*   and Andy Doctorow. Probably wouldn't
*   of been possible without their input.
*  
*  PHP code reduction and minor error
*   checking by Bill Zeller
*
*  This code is released into the public domain.
*/

function shr($a, $b){ // unsigned shift right
    return (2147483648 & $a)?((($a>>1) & 
~2147483648)|0x40000000)>>($b-1):($a>>$b);
}

function mix(&$x) {
  for($a=array(array(13,8,13),array(12,16,5),array(3,10,15)),$i=0;$i<3;$i++)
    for($j=0;$j<3;$j++,$d=$x[0]<<($a[$i][1]))
      
$x[$j]=($x[$j]-$x[($j+1)%3]-$x[($j+2)%3])^($j==1?$d:shr($x[($j+2)%3],$a[$i][$j]));
}

function GoogleCH($url) {
    $url = array_slice(unpack('c*','info:'.$url),0);
    $x = array(0x9E3779B9, 0x9E3779B9, 0xE6359A60);
    for($k=0,$len=$length=sizeof($url);$len>=12;$k+=12,$len-=12){
        $u=array_slice($url,$k);
        for($i=0;$i<4;$i++)
            for($j=0;$j<3;$j++)
                $x[$j] += $u[$j*4+$i]<<(8*$i);
        mix($x);
    }
    $u = array_slice($url,$k);
    for($l=$len,$x[2]+=$length,$i=0;$i<4 && $l<12;$i++)
        for($j=0;$j<3;$j++)
            $x[$j]+=$l>($j*4+$i)?$u[$j*4+$i]<<(8*$i):0;
    mix($x);
    return sprintf('6%u', $x[2]);
}

// [url]http://www.example.com/[/url] - Checksum: 6540747202
$url = $_GET['url'];
echo "url:\t$url\n";
$ch = GoogleCH($url);
echo "ch:\t$ch\n";
$url = 'info:'.urlencode($url);
$link = 
"http://www.google.com/search?client=navclient-auto&ch=6$ch&ie=UTF-8&oe=UTF-8&features=Rank&q=$url";
print("\n\nPage Rank URL: $link");
?>

Convert PHP > Delphi is possible ?

I'm almost finished converting this algorithm to Delphi.
But there is a stupid mistake in my code that I can't find
Probably it is during unaccurate work with data types.
Before last mix(a,b,c) it returns correct b and c but wrong a.
If you wish I can publish my code here.
CU,
uncopyrightable
P.S: sorry for my English.

Windows tool based on above's code:

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

Very good tools ! Thanks !!