[freehaven-dev] an implementation issue

[Roger R Dingledine <arma@mit.edu>]

in terms of communication between the comm module and the haven module:

any given message i get came from another haven module. and hopefully,
the trust db will have been informed of the PK of that haven module
before i get any other sort of transactions from them (meaning, the comm
module should try to remember if it's told me about a guy before, and if
it hasn't build a new message introducing him to me? is this wise or
reasonable?)

but anyway, my immediate concern is this: how do i know where a given
message is from?
i believe we had a slick system worked out for getting binary chunks
(for example, what comes out of the other side of our sign() function)
from one haven to another, but i really don't want to just rely on the
message to declare where it's from. that should be 'out of band' data
that the comm module confirms on its own somehow (or should it?) and
then passes to me along with the message.

or should we ignore this issue entirely, and require all messages to both
a) include the hashed PK of the sender
b) be signed by the sender
this would make each message relatively self-contained...perhaps this is
a good way to do it?

--roger