Google Blogoscoped

Forum

If Only We Knew Google's Secret Ranking Algo  (View post)

Kirby Witmer [PersonRank 10]

Monday, July 31, 2006
13 years ago8,977 views

that's great!!! :)

Nathan [PersonRank 0]

13 years ago #

Lol, that's great!

dpneal [PersonRank 10]

13 years ago #

nice one philipp. very funny!

Pip [PersonRank 8]

13 years ago #

cool algo ;) that's how we perceive it to work :-D

/pd [PersonRank 10]

13 years ago #

hahahah.. this is funny

"// W3C sux! "

Philipp--how could you share me personal info – in public!!!!!

"recipient = "peter[put at-character here]homelandsecurity.gov";"

Tony Ruscoe [PersonRank 10]

13 years ago #

You missed a bit of code right at the end...

// removed by request. 2004-04-24.
// if (owner(url) == "Josh") {
// pagerank = 100;
// }

// added 2004-04-24 – this should be a nice surprise for her...
if (owner(url) == "Diane") {
pagerank = 100;
}

55 Ways readers should understand that... ;-)

alek [PersonRank 10]

13 years ago #


I'm a Perl guy instead of PHP, but I think there is a bug at this line:
   pagerank += countBacklinks(url) / 10000
as I didn't see a limiting upper bound – add this snipper after that:

if (pagerank > 10) {
   if (pageHostedOn(url, 'outer-court.com')) {
   setInternalSecretPageRank(pagerank);
   // Larry says Philipp writes good s**t about us – reward him
   }
   pagerank = 10; // Highest we show externally
}

And I wonder if this was recently added:

if (pageHostedOn(url, 'w3c.org')) {
   setInternalSecretPageRank(-100);
   // W3C sux *and* sells link – penalize those bastards!
}

Seth Finkelstein [PersonRank 10]

13 years ago #

Hee hee.

The comments around the China censorship code might be VERY interesting!

Ludwik Trammer [PersonRank 10]

13 years ago #

Philipp, why don't you relase it under GNU GPL? ;)

Corsin Camichel [PersonRank 10]

13 years ago #

Does not compile!

Haochi [PersonRank 10]

13 years ago #

if (inString(content, "tiananmen")) {
   if(pagerank > 6){
   pagerank -= 6;
   }else{
   pagerank = 0;
   }
}

This is an addition to Google China's algorithm.

mak [PersonRank 5]

13 years ago #

Perfectly done! as always.
I wonder how can you come with something that special everyday!!

keep surprising us ;)

J. McNair [PersonRank 10]

13 years ago #

Other things you may be missing:

// Larry <3 Marissa 4Ever 2004
if (secretSemanticAlgorithm (content, "Marissa Meyer") == ROCKS) {
   pagerank *= 2;
   addtoInternalSecretPageRank(500);
}

// Best project at the best company ever – NW
if (secretSemanticAlgorithm (content, "Niniane Wong") == ROCKS) {
   pagerank = ceil (pagerank * 1.5);
   addtoInternalSecretPageRank (200);
}

// NOTE: Don't tell W3C about Semantic Algorithm, it might
// cut into Adwords sales – Sergey

Piotr Konieczny [PersonRank 9]

13 years ago #

Nice work Philipp!
Now, let's rewrite it into other programming languages :-)

stefan2904 [PersonRank 10]

13 years ago #

lol. great idea Alek :p

Philipp Lenssen [PersonRank 10]

13 years ago #

> I wonder how can you come with something that special everyday!!

Thanks, blame this one on Alek :)
komar.org

alek [PersonRank 10]

13 years ago #

Philipp, as typical, is too generous – I sketched a rough idea, but as with most things, implementation is key ... and it's his PHP code – well done as always Mr. Blogoscoped.

Just be glad I didn't write it – my Perl code is often not even understandable by me! ;-)

Haochi [PersonRank 10]

13 years ago #

Just translated half of the script into PHP, continue tomorrow, crack the "Google Algo", haha ;)-
haochi.chen.googlepages.com/ch ...

Philipp Lenssen [PersonRank 10]

13 years ago #

Heh. Funny.

Philipp Lenssen [PersonRank 10]

13 years ago #

Is "XhtmlValidator.php" for real?

Haochi [PersonRank 10]

13 years ago #

Yes, no kidding.
bermi.org/downloads/XhtmlValid ...

Sohil [PersonRank 10]

13 years ago #

blacklist2 = getList('c:larry-page-hatelist.txt')

That's Beautiful.

Yo Lar Am I on the Hatelist.

CJ Millisock [PersonRank 10]

13 years ago #

Awesome! lol

Josue R. [PersonRank 10]

13 years ago #

hilarious!

Tiago Serafim [PersonRank 4]

13 years ago #

Nice post Philipp!!!

digg for it: digg.com/software/If_Only_We_K ...

Suresh S [PersonRank 10]

13 years ago #

Bool Gbot=0N;

if (pageSearch(url) == "Microsoft") {
Gbot=OFF;
}
else {
Gbot=ON;
}

Personman [PersonRank 8]

13 years ago #

Awesomely awesome. Good job!

Tadeusz Szewczyk [PersonRank 10]

13 years ago #

What this post highlights to me is: The urgent need for an open source search engine that really works and is really popular.

Why didn't all the open source search engine projects didn't take of yet? Everywhere else there is a strong open source alternative where there is a dangerous monopoly: With operating systems, there is Linux, with browsers there is Firefox, with graphic editors there is Gimp(shop) and so on and so forth...

Nutch, Grub and so on are no real alternatives yet..

Suresh S [PersonRank 10]

13 years ago #

Yeah as Tadeusz Szewczyk said:"We need open source search Engine" "Open Source AdMarket".so that whatever earned by AdMarket can be dedicated to Open Source Projects.

Dannie Jost [PersonRank 0]

13 years ago #

Actually the secret is know. There is a whole bunch of competing monkeys fiddling with the rank engines day in and day out! Code monkeys?

Haochi [PersonRank 10]

13 years ago #

Dannie, I think you missed something, it's pigeons. ;)
google.com/technology/pigeonra ...

NateDawg [PersonRank 10]

13 years ago #

I think Haohi nailed it :D

Seth Finkelstein [PersonRank 10]

13 years ago #

The problem with an Open Source engine is not the source as much as the CPU – that is, the server-power required to run everything for a service would have to be donated. Linux, Firefox, Gimp, etc are all run *locally*.

Ilya Kniazeu [PersonRank 1]

13 years ago #

Thanks, Philipp! Real fun to read.

Alexandre [PersonRank 1]

13 years ago #

Very funny! It is like coca cola secret formula. If you kwnow Google must kill you!

Philipp Lenssen [PersonRank 10]

13 years ago #

Maybe that open source search engine would need to be distributed among many different computers. A small program that, when the PC is idle, computes queries in the background. A bonus would be that you couldn't censor this easily, because it's not controlled by a single source.

Haochi [PersonRank 10]

13 years ago #

Just finished translating the script, here it is, play with it if you like.
googlified.com/junks/check.php ...

Suresh S [PersonRank 10]

13 years ago #

global PgRank;
if { [catch foundPage] msg } {
   foreach word (getKeyPhrase()){
   if { [regexp -nocase $word "Don't be evil"] && getTitle(url)!="Google" } {
   PgRank=0;
   } else {
   PgRank++;
   }
   } else {
   response.redirect(" google.com/ig\");
   }

Philipp Lenssen [PersonRank 10]

13 years ago #

Haochi:
> $pagerank += countBacklinks($url) / 10000; Help me with this!

Can't you just screenscrape the page count from a Google search on [link:thisdomain.com]?

Haochi [PersonRank 10]

13 years ago #

I thought it would be easier to do it with Google Search API, but I have no experience with the API.

Seth Finkelstein [PersonRank 10]

13 years ago #

It would be very, very, difficult to have a distributed database and computation engine. In fact, that's what Google does, and it's why they're building huge server farms near hydroelectric power plants.

gabel [PersonRank 0]

13 years ago #

awesome!!

Erik [PersonRank 0]

13 years ago #

// added for fun
if ($TitleOfPage($url) == $search && $internalPageRank<rand(0,10)) {
setHiddenPagerank =-100;
echo "Did you mean: ".opposite($search)."?";
}

This thread is locked as it's old... but you can create a new thread in the forum. 

Forum home

Advertisement

 
Blog  |  Forum     more >> Archive | Feed | Google's blogs | About
Advertisement

 

This site unofficially covers Google™ and more with some rights reserved. Join our forum!