On Mon, 2 Dec 2002, Peter Fairbrother wrote:
What I don't understand is how a node knows the location of a person who moves about in the first place.
The node spans a cell. Similiar to your cellular phone, you can link an ID to a cell. Within the cell you can use relativistic ping and/or signal strength (that's how mobile phone localization is done today). Since cells overlap you've got a lot of constraints to get a position fix.
Also, I don't like the idea that my location is known by the location of my equipment. But I know very little about geographical routing.
Your location is already known, whether you're using wire or wireless. Wireless has limited range, cables are expensive enough so that their lenght is being minimized. Traceroutes and signal pings and already existing IP location databases make anonymity a myth. The only way to address it is to use anonymizing proxies/traffic remixing. Geographic routing is intrinsically resistant to address spoofing (neighbours will refuse routing packets from obviously bogus origin). If you want to avoid disclosing your physical location, use a higher, anonymizing protocol layer.