Yes, but in the vast majority of (unreported) cases, what gets tracked by
law enforcement is everything from Nigerian scam artists (which gmail has a
huge, huge infestation of.. I know for a fact - attributable to him, but not
in the media - that chris lewis of Nortel has blocked gmail, entirely.. if
you know Chris as I do, that's saying a lot)
RIAA etc subpoenas are civil subpoenas, and are handled with warning given
to the user who is the subject of the subpoena etc - plenty of procedures to
protect your users as well. Unfortunately, they, and the jackass variety of
subpoena (X defamed Sonia Gandhi) etc are what frequently get cited when
ruling IPs PII in a way that effectively makes them out of bounds for ISPs
to store them.
srs
> -----Original Message-----
> From: [EMAIL PROTECTED]
> [mailto:[EMAIL PROTECTED] On Behalf
> Of Rishab Aiyer Ghosh
> Sent: Thursday, July 10, 2008 11:41 AM
> To: [email protected]
> Subject: Re: [silk] google email anonymity
>
> On Wed, 2008-07-09 at 19:04 -0700, Suresh Ramasubramanian wrote:
> > There have been a few fun cases where RIAA subpoenas have been sent
> to IPs
> > that were actually permanently assigned to a university laser printer
> > (duke? tufts?)
>
> yeah, but also lots of cases where law enforcement used IP addresses to
> track down individuals who may or may not have done anything wrong.
> well, of course law enforcement can do that, but if something is useful
> for them as having a high chance of identifying an individual, it's
> useful to anyone else too. the argument that theoretically, IPs could
> be
> non-personal is made frequently by ISPs but for some reason ISPs just
> don't provide the facts to back that up, which should be easy for them,
> more than anyone else.
>
> what are the probabilities that when X logs in repeatedly as
> [EMAIL PROTECTED] from 2 or 3 specific IP addresses over a few
> weeks that you can identify X as the person with access to that email
> address? and that you can identify the address X accessed the email
> address from? i'd say the chances are pretty high, and the ISPs have
> the
> data that _could_ suggest otherwise - if in fact IPs are as useless as
> identifying people as they claim.
>
> -rishab
>
>
