$Id: global-address-privacy-leak-fear-and-uncertainity.html,v 1.4 2007/07/16 13:58:48 itojun Exp $That is not true. The truth is, it is very very difficult to keep your privacy when you try to get information from someone, or shop from some website. You have to be like James Bond to achieve this. Not everyone has Q branch to support them, not everyone can attract bond girls like James, and not everyone likes vodka matini "shaken not stirred".
Anyways.
Regardless from the type of IP address (global or private), use/non-use of NAT, or IPv4 or IPv6, your identity is keep tracked by HTTP cookie. HTTP cookie allows web shops to keep your shopping cart by identifying your browser. But if the shop correlates the HTTP cookie with the transaction (of course the shop is doing it always), your browser and your credit card information will easily be associated. If you are in the US or Korea, Social Security Number can be associated as well.
Additionally, your operating system can easily be guessed by the use of "passive OS fingerprinting" technology. Except with very specific operating systems and/or NAT boxes, the shop, or intermediate boxes, can sniff your packets, look at the "signatures" in TCP header, and identify your operating system.
Moreover, if you use Microsoft NetBIOS protocol, almost every personal information on your system is visible by default in the past Microsoft operating systems! (I'm not too sure about the details, I'm an UNIX guy) I hope this is disabled by default someday, but if my memory serves it has been enabled by default for some time.
With IPv6, there's a document called "Privacy Extensions for Stateless Address Autoconfiguration in IPv6" (RFC3041). However, in my humble (or not so humble) opinion, this document is a way to invite a lot of security holes. Since your identity is tracked no matter what you do, you only add complexity within your OS kernel (IPv6 stack) to buy nothing, or to buy a false sense of security. So you only get certain amount of additional complexity, which is of course additional amount of code, which is a possible source of bugs (more code, more likelyhood of bugs) and additional memory footprint (which is a problem for embedded devices). See KISS principle.