no circuit loops?

[Roger Dingledine]

On Thu, Oct 23, 2003 at 05:27:26PM -0400, Paul Syverson wrote:
> But that's just a default yes? And, it's also just temporary. We

Actually, the default is 2+n, where n is chosen from a geometric
distribution around the config option CoinWeight (which really needs a
new name). CoinWeight is generally .01 these days. Is there any known
value to choosing path length this way against the adversary we expect,
or should we simplify?

> still need to consider whether we get more vulnerability from
> allowing loops or disallowing loops (or e.g. disallowing ABA loops
> but allowing larger ones). And we should probably work that out before we
> have route selection code that is new and shiny.

Actually, I should make route selection code that can do it any of those
ways and still be shiny. We will see. :)

But you're right, we should solve this. At the least it should go as an
open research problem in the design paper. If you have more of an answer
than "gosh, hard problem" then let's hear it. :)

--Roger