home
How would you recognize it instantly ? (none / 0) (#7)
by Joliphant on Sun Feb 17, 2008 at 06:38:02 PM EST
You have a system dumping data onto a disk or flash memory just how do you recognize it instantly ?

Or someone hands you a database or a set of files just how do you divine their content instantly ?

Making a lot of mistakes ?

Just where would you place their percentage ? This is the only incident I have heard of in the past year. I can't quantify the rate because I have no idea how many times they undertake this kind of operation or how many times we hear about the result.

[ Parent ]

For starters (none / 0) (#8)
by badger on Sun Feb 17, 2008 at 06:56:19 PM EST
email is in text format - at least the headers are. You can just view it. If you couldn't view it there wouldn't be much point in collecting it.

To deliver it to an account requires information specified in the RFC for mail - things like a 'To' address as well as a 'Deliver-To' header, so I can either look at a random sample of messages or write a simple script to check every message. Or something like sendmail can be used to validate it too, although it's easily scripted in Python or Perl (having done it a number of times).

In fact if I don't do something like that, I have no idea if I'm collecting the mail I want to examine at all - mistakes can omit things just as easily as include hundreds of wrong addresses.

Delivering mail to the correct address is not some unfathomable, error-prone process - how often do you receive email intended for someone else's account?

So assuming they were smart to enough to check that they were receiving the mail they were interested in, they had to know they were receiving lots of other mail they weren't authorized to receive. There's no way around that that I can see.


[ Parent ]

That makes assumptions (none / 0) (#9)
by Joliphant on Sun Feb 17, 2008 at 07:08:03 PM EST
without facts in evidence to support them.

  1. The email is being intercepted as text streams at time of transmission and not being retrieved as extracts from an exchange server or other email server.

  2. If they are being intercepted in real time you assume they have they have remote monitoring of the device used to do the interception. Simply mistyping a wildcard can throw your intercept filter way off.

I don't want to drill down into endless detail on this. So lets cut to the chase.

Your argument rests on two pillars. One, that software and I.T. operations are being conducted with a very low error rate. That people don't make frequent mistakes with very complex equipment and that these mistakes may not persist until subsequent review.

Experience argues otherwise.

The second pillar is that this happens often. You haven't presented anything into evidence that supports that argument.

[ Parent ]

The assumption it makes (none / 0) (#10)
by badger on Sun Feb 17, 2008 at 07:56:00 PM EST
is that everyone is receiving RFC-compliant email. If that's the case, then once the message leaves the sender, anywhere in the datastream you access the message the RFC-envelope will contain the necessary information to differentiate between the suspect and anyone else.

It isn't necessary to drill-down any farther than knowing how an email message is structured, routed and processed by mail transfer agents. To screw up the way they did you either have to a) set things up with absolutely no concern for the privacy rights of others on the same network or, b) do it intentionally.

But to strectch a point, I suppose someone who is incredibly lazy could collect all of that mail by accident, in which case he'd be lucky to collect or find the mail he's actually interested in. I'm not sure which is scarier - the FBI being that invasive or that incompetent, but there is no third choice.

As to the frequency of 'errors', there is a report from the DOJ itself (I believe) that indicates how many times the FBI and others have violated the limits of the law and legitimate warrants. If you really aren't aware of that, I'll try to look it up for you.


[ Parent ]

No (none / 0) (#11)
by Joliphant on Sun Feb 17, 2008 at 08:18:09 PM EST
You are making assumptions again.

  1. That the mail is RFC compliant
  2. All the data is passing through a central point in text format accessible to the FBI.

Both are just not so as often as not.

Remember unless its an ISP hosted email, you will want to intercept internal mail to the address as well. In which case you may have to deal with anything from groupwise to notes. There is also the matter of companies that use secure or otherwise encrypted email systems.

As I stated earlier your argument rests on the idea that its possible to provide a very high level of accuracy in collecting this information. Given the arbitrary variance in I.T. environments this will not always be the case. It all boils down to the error rates.


[ Parent ]

  • Premium Ads

  • Blog Ads

  • Contribute To TalkLeft

    donate to TalkLeft