...

Assume we create the alt.anonremailer.net newsgroup mechanism that Jonathan Rochkind recently suggested (and it worked). > > Could we then not use the newsgroup, in combination with a net of > well-known remailers, to give us the capability to have some remailers at > unknown locations by having some remailers post encrypted reply blocks as > their "addresses"?

This is just painfully non-scalable. Sure it will work for now, but its not something that will last once large numbers of people begin using it.

Why? Which part, the whole idea of a remailer control newsgroup, or just the idea of remailers with unknown locations? I'm not sure how reliable remailers with unknown locatoins would be (one remailer in the chain goes down, your unknown remailer can't be contacted, and there's no easy way to verify whether the chain is still intact any more), but I'm not sure I actually see anything non-scalable about it. Nor about the "alt.anonremailer" concept. If you've got 500 remailers posting once a day, your still not the largest newsgroup out there. And when combined with a realtime verification system (you get the address of the remailer from the newsgroup, and _then_ you connect to a certain port and get an acknowledgement message, to make sure the remailer is up), you could lower traffic yet further. Maybe you only need to post once every ten days with the "my remailer is here" message. This could then accomodate some 3000 remailers and still not be as high traffic as comp.sys.mac.hardware. And I think by the time there are 3000 remailers to be accomodated, the net will be easily handling full motion video, and people will be exchanging 20 minute long quicktime movies in the newsgroups, and we wont' have a bandwith problem at all. But maybe I'm missing something obvious. Enlighten me.