1994-04-06 - Re: Proposal: some more standard remailer features

Header Data

From: michael shiplett <michael.shiplett@umich.edu>
To: cypherpunks@toad.com
Message Hash: 00eea244aae4fd458494eab173745751ec20333ae66056f7f9235b7a7767158e
Message ID: <199404060505.BAA08693@totalrecall.rs.itd.umich.edu>
Reply To: <9404060309.AA25086@geech.gnu.ai.mit.edu>
UTC Datetime: 1994-04-06 05:05:53 UTC
Raw Date: Tue, 5 Apr 94 22:05:53 PDT

Raw message

From: michael shiplett <michael.shiplett@umich.edu>
Date: Tue, 5 Apr 94 22:05:53 PDT
To: cypherpunks@toad.com
Subject: Re: Proposal: some more standard remailer features
In-Reply-To: <9404060309.AA25086@geech.gnu.ai.mit.edu>
Message-ID: <199404060505.BAA08693@totalrecall.rs.itd.umich.edu>
MIME-Version: 1.0
Content-Type: text/plain


"r" == Ray  <rjc@gnu.ai.mit.edu> writes:

r>    Here are some proposed remailer standards some of which I have
r> already implemented.

[ command formatting section deleted ]

r> Anonymous Posting:
r>    On any mailer which supports virtual addresses, the following special
r> feature shall be implemented:

r>    If the virtual address contains any '.' characters, the address
r> is first assumed to be a newsgroup. If the newsgroup exists and/or it
r> is not blocked by the operator, two possible actions can take place.
[ details on newsgroup posting deleted ]
r> Example:
r> ::to remailer1#remailer2#talk.politics.crypto

r>   If asked, I will supply the magic perl subroutine needed to do
r> this.

r> [note above, I have eliminated the redundant "request-remailing-to".
r> When mailing through a remailer, you know the mail is going to be
r> remailed. ::to is easier to type]
  I suggest changing "to" to the previously mentined
"post-to"/"send-to" convention. This eliminates the need to perform
parsing magic on the virtual address. Also it's a simple issue, but
what's the syntax for defining a variable, e.g., NNTPSERVER or
NEWSGATE?

[ details on fragmented messages deleted ]
r> Now when ann's remailer receives a two parted message, it queues
r> each piece until it gets the full message (timing out after a few
r> days) After all pieces are received, it removes the envelopes,
r> pieces the message together, and sends the message off to darkmodem
r> (which may be a virtual address for lightmodem#bob's_remailer)
  Sounds like a nice feature.

r> Additional ideas:

r>  A command ::error-to to specify where errors encountered during
r> processing of the message should be sent. e.g.
r> ::error-to idstring an99999@anon.penet.fi 
r> or
r> ::error-to idstring alt-waste@cs.utexas.edu

r> [idstring will let you know which message the error was for]
  Another good idea, but how would I, as a user, know with which
idstring one of my messages is associated?

r>  I also propose ::route which would specify preferences preferred
r> for remailers when searching for other remailers to chain your
r> message to. e.g.
r> ::route Private
r> [attempt to chain to remailers which are running on single-user
r> non-public machines first]
  I've followed the arguments for having the remailers keep track of
each other's availability. This is fine as long as one can strongly
trust at least one of the remailers. The chaining functionality also
belongs in the mail client--even more so than in the remailers.  With
extensible mail environments, e.g., mh/mh-e, this should be possible
without too much difficulty.

  I don't know if it's been suggested, but has anyone created a
remailer that scans a newsgroup for posts addressed to it in some
manner, e.g., an X-header or the first non-blank line, and then
handles the post as if it had received it via mail? Sort of a Kibo
mail gateway.

michael





Thread