One of the most enduring myths of the internet is that it was designed to survive a nuclear war. If NORAD gets obliterated, the system as a whole keeps running. This is true for many of the systems of the internet but anyone who administers Apache, MySQL or SVN server can tell you that this isn’t as simple as it sounds.
One of the worries of the internet is that we all depend on google too much and that search is too important for one company to handle. I try to use all the search engines out there in rotation (although google always seems to find what I’m looking for the best).
There are a number of projects that are active at the moment to build an open, distributed search engine. For example, Grub, Faroo and YaCy. I like the idea of these but have a few concerns. What’s to stop the client, running on your computer in the background, downloading a lot of content that’s objectionable or illegal? Perhaps we should all be running such software as an act of civil disobedience to make it impossible for the police to track traffic.
There are other sorts of search that could be distributed. Image recognition is very processor intensive but should be parallelisable.
A distributed clone of the ESP Game could be written. Not that I want to knock the existing version of the game, it’s great but centralised. User’s give google all this data. Are they forced to give it all back to the users or just what they want to show.
Is it really that important that search is distributed? Eventually, I hope that it becomes that way. It’s a similar deal to Windows vs Linux or Java vs PHP or Wordpress vs Blogspot. The cathedral vs the bazaar.
