I find it a very interesting piece of cultural difference that US citizens don’t care about their data being traded and (mis-) used by virtually everyone who can afford to pay for it. Why in the world should I trust a commercial company, whos first goal is to make money, more than the government which should be in my service (not that I trust them, either)?

The “update my data” problem is significantly harder, for the reason I stated before: the entirety of your personal information is spread out amongst vast numbers of gigantic log files. These files are stored on GFS, which is not designed for efficient random write access. The scale problems are tremendous.

This is not to say it can’t be done, only that it’s hard to do. It would be great if one of the major search providers implemented John’s suggestions. My comments are in response to the John’s tone of voice, through which he suggests that this problem is easy to solve. I don’t believe that is the case.

Policy issues aside, it’s a technically hard problem due to the sheer scale of the infrastructure involved in collecting, organizing, crunching, and using that data. There are home-grown tools, third party commerical tools we don’t have souruce code to, and many groups involved.

Had the system been designed with that goal in mind from day #1 (who’d have thought to do that years ago?), it’d be one thing. But this is a very large retrofit project. And we’re still strugling to find enough great engineers to build things we’ve had on the drawing boards for quite some time now.

The same people who seem to worry that GW will toss them into Guantanamo after viewing their search history have few concerns about getting quietly and secretly assaulted by an SE marketing department gone wild.

Processing search data via a control panel for those that request this should NOT be a huge challenge (this is what cheap scalable parallel processing was made for, OR just use a local program on the users computer!).

Google Analytics (formerly Urchin) does a far more robust analysis of websites at a far greater level of detail than would be needed to simply review search queries.

It’s an issue that Paul Martino and I at Aggregate Knowledge have been following closely as we have been building out our attention/behavior based recommendation engine. We must give users access to that data and allow them to really have control over it. We aren’t there yet but we are actively working with Seth Goldstein at Attention Trust to figure out the best way for us to do it. I envision a future where you’ll be able to use an Attention Trust interface to get at data that you have created on our servers or anyone’s.

One place where I think that you and others aren’t going far enough is in only asking our online service providers to give us this control. Why doesn’t Visa give me this control? How about my bank? I’m just as concerned if not more so of THAT information getting into the wrong hands than I am about my search behavior.

I’m not sure, to be honest. The idea of someone reading your mail when you leave yourself logged on at a public terminal is bad enough – imagine if they could see your search history, the times your most likely to be online, links clicked on, etc … and then make changes by filling out a form.

Anyway, very, VERY, complex. Desirable, but almost impossibly hard to do right.