The search engine and advertisement agency Google may know you better than most of your friends.

Imagine... a world with a wonderful search engine called g∞gle :

G∞gle : you got us, don't wriggle

g∞gle is not evil ; what happens when you use such an open service ?

• the g∞gle website
  First of all, G∞gle knows what you are looking for when you use the G∞gle engine, even from an external website. Thus, each search is systematically linked with the originating IP. Moreover, if not spoofed, the http_referer reveals where you are coming from. Now, one important point is that as a power user © you're maybe using G∞gle's preferences, for example to display 100 results instead of 10. That means that you have the luck to store on your computer a cookie^[1] with a unique id ; that allows G∞gle to associate whatever you do on the website (and more, cf below) with this cookie, even if you use different ips (dynamic attribution, proxies, tunnels etc).


• g∞gle's translator/maps/earth/answers/fr∞gle/news/groups/blogs/etc
  Naturally, if you are exploiting these features, G∞gle knows what you buy, what you want to translate, which communities you are in, what kind of information you're interested in, where you may live and go, and so on.


• g∞gle scholar/code
  If you're working on a confidential project, what a good idea to take advantage of these tools to obtain some academic information for it.


• g∞gle talk / g∞gle mail
  For backup purposes, your conversations and private/professional e-mails are kept in the G∞gle database, as well as the names and addresses of your contacts. So if you lose data G∞gle may retrieve it for you, cool. This topic concerns also people who send mail to your account : their e-mails will be scanned (unfortunately without explicit consent) for advertising purposes.


• the adword program
  If g∞glesyndication.com is not in your /etc/hosts, or something similar, you have the right to see the ads on websites using Adword. That is, you inform G∞gle of what you are currently visiting without doing any kind of search ; G∞gle finds for you the relevant ads to display.


• the pagerank bar and web accelerator
  In this world full of matrices of yellow dots, it's cool to see in real time the rank of the site you're visiting. The price to pay is just to ask G∞gle "hi, i'm visiting this very webpage, is it g-popular ?" Same holds for prefetching pages : convenient but not really private.


• g∞gle's analytics
  This one is a little bit insidious because it is completely transparent for the end-user. It is a very well thought tracking system and analyzer, building statistics for webmasters.


• g∞gle's video
  Viewing an embedded content requires an indirect connection to G∞gle's databases, giving information like what is the website you are on, what is the media you are looking at, etc.


• the desktop
  To help you to quickly search your files from anyway, G∞gle offers you to upload a listing of your personal files. Are they now that personal ?


• the apis
  The G∞gle technology will be (is ?) embedded into softwares, mobile phones, etc, and will provide relevant links for some resources without that the customer knows that he is using the underlying services of G∞gle's servers.

These details mixed together gives G∞gle an ultra combo (the best one in Killer Instinct iirc) so the search engine can typically know a little bit who you are.^[2]

Nothing is wrong with this kind of profiling ; it is pleasant to get such a kind of customized service. However it is a good thing to be lucid about how all this is done. With a little bit of data mining, some very interesting correspondences can be extracted from databases of requests. Databases do not forget. Of course it necessitates huge storage space and awesome computing power ; but is that really the problem ?

A potential method to avoid such exposure with G∞gle would be to generate artificial requests flooding and camouflaging the real ones. Anyway, lets return to Google now. A clever compagny.

Eventually, if you read Google's privacy policy, you realize that your personal information won't be sold... Everything is automatic, it is Deus Ex' Helios. No need to worry. So far so good.

[1] In the spirit of Bonuscards swaps, someone could write a program allowing people to share and exchange G∞gle cookies with the same preferences.
[2] Of course all these considerations also apply to your ISP, or to anyone sniffing your line ; TCP/IP communications are intrinsically spywary.

Little Neo, oct 2005 / upd feb 2007

zZz

Home