Douwe Osinga's Blog: Why we should leave Google behind

Monday, September 8, 2003

Why we should leave Google behind

Google has been a good friend for over 5 years now. It has taken the Internet world by storm. They have build a brand purely on the quality of their product, almost no advertising was involved. In many ways, Google brought back the sensibility that had slipped the Internet during the boom. The clear user interface, the seperation between advertising and commercials, even the humor, it was all very much the way Internet was supposed to be. Lately, stories around Google sound like the end of a beautiful friendship.

A year or so ago, the Scientology Church threadened Google with legal actions, if Google would not remove a link to, some Norwegians with an alternate vision on Scientology. Google complied. Since some time, xenu is findable again through Google again, though. However, last month or so, Google removed all references to Kazaa lite from its database, because Kazaa (the irony) had complained about copyright infringements.

Google, of course, is a company and has to think about shareholders value. However, when Google starts to filter its results per requests of governments and removes links that are not to the liking of a certain government, we can no longer trust Google with the central position it currently has in our information eco-system.

It is not only that. Just like we shouldn't depent on one company for something important as an all present OS, we can't really depent on one search engine that decides with secret algorithmes what page is the most relevant. Our freedom would be much enhanced if we would know how search engine results where produced, even more so if we could decide which algorithme to use.

What is needed here, is a seperation of data and algorithmes. The webcrawl programs that scourge the Internet for fresh content and are not that different between search engines. Some are faster maybe, but in the end they build up a catalog of what there is on the Internet. The collecting of all this data could be done by a widely used p2p program (like loomsmart's grub does). This information should be free to use and free to inspect by all - it is after all just another view of the World Wide Web itself.

On top of this distributed database, anybody should be able to build search engine algorithmes. Some will imitate Google and put pages that are linked to on top. Others could do a more intrinsic analysis of the content of the pages. On top of it all a variety of search engines could be build and no doubt a lot of other applications too. Just have a look at the results of the recent Google competition.