The problem with publicly revealing every aspect of the search engine algorithm is that the people who want to pump their stocks, sell you their junk, and enhance the size of your sex organs want their pages to appear on top. If the Google algorithm were completely open, Google would be much less useful, as every search you do would have the unethical scam artists filling the first page. Searches for a company's product might take you only to pages plugging the company's competitor, or even to phony pages put out by someone trying to steal money from you.

Now, maybe I'm wrong and people can come up with an algorithm that can be public but can't be exploited. At best, it would require continuing modifications, which might mean that pages you can find on one day can't be found on the next.

I totally disagree. A truly well-constructed algorithm should not be easily exploitable even if all the details are known. Security through obscurity is no security at all. Yours is one of the most bogus arguments frequently made by proponents of closed-source software. All the source code of the Linux operating system is publicly available, yet Linux is several orders of magnitude more secure than Microsoft's closed-source operating system.

Just because you 'know' how rankings are derived doesn't mean that they're immune from distortion or hype. And when the results come out more or less the same as Google, since that's what Jimmy is talking about, then what do you do?

Dear Salon,

This ws a nice article, but please - why does no one pay attention to the disaster that Mr. Wales created, and which he still holds total control over? Wikipedia is a mess, and it is clear evidence that he can't manage, can't delegate and can't organize problem solving. He's taking all the fame of Wikipedia, and is raising investment capital, while Wikipedia is a complete disaster. There are 7 paid employees at Wikipedia, and its great taht he can get all the free labor, but there is no oversight, and the content has suffered. Most administrators on Wikipedia are under the age of 18, and they have no training. You'll find 10th graders banning university professors, or teenagers who taunt private people by retaining biographies that aren't warranted. Wales just lets most of this happen - he doesn't care.

Wikipeida has a budget of about half a million. On its shaky reputation, Wales has raised over 6 million dollars in venture capital. Small peanuts in the IT world, but he's already shown that he can't manage. He's already shown that he has no moral fiber - when he kept the 24 year old Essjay (Ryan Jordan) who'd lied ot the New Yorker about (not really) having two PhDs. Wales backed him up until the public outcry became to loud. That's Wales - the random manager. He took that decision by consensus (the consensus being that he made poor decisions).

Most importantly, Wikia has no real product - Wikia is the ultimate dot.com - money poured into a hopeful concept. He wants to compete with Google, because he swiped Sangers concept and made it work - sort of.

I wish your writers would take a good hard look at Wikipedia, and how dangerous and dysfunctional the product is - which is only part of the reason the content is unreliable.

If you want the real inside story on Wikipedia, come read the Wikipedia Review. Wales calls it an "attack site", mostly because he can ban criticism on Wikipedia. Wikipedia Review discovered the Essjay Scandal, and they still hold the pulse of Wikipedia (which is fading fast). I'd recommend Wikipedia Review to any potential Wikia investors - you can read what Wales is really good at - and not good at.

Thanks

"Not Impressed with Wikipedia or Wales" at http://www.wikipediareview.com

The real challenge for a search engine is based on the recognition that different users have different search intentions and priorities. Unless a search engine is equipped with the ability of the individual user to modify the rules which control both the selection and ranking of returns, it falls short of the ideal. Open source may merely enable a searcher to perceive why the engine is not returning the selections (s)he wishes but not enable the searcher to do anything about it. The flexibility of the search engine is the key if we expect ultimately to push Google to change.

I love this part:

"Search is broken, for the same reason all proprietary software is broken"

Looks like GRUB is working fine though:

(Must be my browser and OS)

Warning: mysql_connect() [function.mysql-connect]: Unknown MySQL server host 'sv-grubdb1i.looksmart.com' (1) in /opt/grubwww/home/services/urlstats.php on line 45

Warning: mysql_select_db(): supplied argument is not a valid MySQL-Link resource in /opt/grubwww/home/services/urlstats.php on line 46

Warning: mysql_query() [function.mysql-query]: Access denied for user 'nobody'@'localhost' (using password: NO) in /opt/grubwww/home/services/urlstats.php on line 49

Warning: mysql_query() [function.mysql-query]: A link to the server could not be established in /opt/grubwww/home/services/urlstats.php on line 49

Warning: mysql_fetch_array(): supplied argument is not a valid MySQL result resource in /opt/grubwww/home/services/urlstats.php on line 50

IndexedUrlInfo URL not found - URL may not be in the database, or may not be in a normalized form.

Open source operating systems can be secure because there is some data (passwords, crypto keys) that is not public. Likewise, an open-source search engine might be secured from "attackers" if certain data, like the relative weights assigned to various page features, are kept secret. I run Linux as well, and am composing this message on a Linux system. On the other hand, Google's search engine is one of the world's biggest Linux apps.

I'm not saying that it is impossible to have a totally open search algorithm that can't be spoofed by consultants promising higher page ranks. I'm only saying that it has never been done.