We're not trolling your personal life....really...its just metadata
By pmonday on May 11, 2006
Information is power, but what is information about information (metadata), is it powerful too?
Of course it is, but its a different kind of power and its usefulness (or abusefulness) is, of course, all in the eyes of the owner of the metadata.
Let's say, hypothetically, that you knew that I:
What could you deduce about me? Have you crossed the line into my private life? Could you use the information to cross the line? I'll only answer the last question, the answer is emphatically "yes".
Let's say, hypothetically, that I placed
- 15 calls to Jonathan Schwartz today
- 30 calls to Jonathan Schwartz yesterday
- 10 calls to Jonathan Schwartz the day before
You also knew that none of these calls were answered and that the rapidity with which the call went unanswered increased (started out with 6 rings and by day 3 it was 1/2 second). What can you deduce about me? Have you crossed the line into my private life? Could you use the information against me? Again, with the last question, the answer is emphatically "yes".
I'm a guy that doesn't really have that much to hide, at least I don't think I do. But I have to tell you, I think you should stop to think about today's news and figure out how it applies to you, not the politics of it, not the emotions of it, but the raw lesson about metadata and what a person can learn simply by knowing the outline of your world rather than knowing the details of the data.
Metadata provides a framework and navigation technique for the information that lives within the model...I know I'm taking liberties here...metadata is a lot of things to a lot of people (its even two words sometimes, meta data). Further, its "depth" could easily be argued about depending on our domain, is my height and weight a part of my metadata or is this actual information. Let's say, for the sake of argument, that our domain is, well, knowing things about people. The contents of a phone call is data. The time of the call, the participants, and the length of that call are metadata about it. Let's say I collect information from 5 people over the course of a week and then extend the metadata collection from those 5 people to the top 5 calls on the list for the next week and so on and so on for a year.
Then, I take the information, toss it at a grid (you know that I'm a Subaru owner with an REI credit card so you can likely deduce that I probably don't have a big data center in my basement) and derive the connections in the grid. Not only can you quickly deduce who my immediate social network is, but you can also quickly deduce the probability of my coming into contact with a member of the extended social network. Using times and lengths of calls, you could probably even get a better accuracy of that probability. Toss in metadata about the location of the phone calls (quite easy to do with cell phones) and...well...
There are lessons here about your computing environment as well. How much information are you comfortable about people knowing about your computing environment? I would hope that your answer is "that depends on who I'm giving the information to". Our local bookstore, the Tattered Cover, fought for your right to privacy of the record of your purchases from the bookstore. To their credit, Google went to bat as well, whereas other engines have not. Keep in mind this is all METADATA this is not the actual data that you browsed...