Auto-Verifying of Sigs

-----BEGIN PGP SIGNED MESSAGE-----

Thinking about this requiring/checking sigs thing, I thought of 
something...

Really, the only "unknown" with signed messages is whether they are valid 
or not; it's pretty easy to distinguish the unsigned posts.  Furthermore, 
it seems to be my observation with verifying digsigs (as I do in 
non-crypto groups I subscribe to) that the vast majority of sigs will 
turn up OK.  It seems, therefore, that expending a lot of effort to 
change the current list to allow this would be wasteful considering the 
relatively few times that it would produce any useful information.

May I propose a "better" way (you be the judge here): Proxy the job.

Have a 'bot subscribe to the list (through whatever way), armed with a 
complete keyserver keyring.  Its only function is to check all signed 
messages from the list.  Unsigned messages, messages with sigs that 
checked OK, and messages signed with unknown keys would generate no 
response from the 'bot.  A failed sig, however, would cause the 'bot to 
send a (digitally signed, optionally) message to the list to the effect 
of "This message here didn't check OK" (complete with disclaimers and 
warnings about trusting authorities blindly).

This would be a totally automated way of checking sigs, and wouldn't 
involve any new code on the list's part.  Those who didn't want the 
intruding messages could killfile the 'bot, and the rest of us wouldn't 
be bothered with redundant information on every post.

What say ye all?  I can tentatively volunteer my business account to do 
the work (have to talk to my boss about it first, as that account has to 
pay for volume and phone time).  I'll play with some code in the meantime 
and see what I can come up with.

-----BEGIN PGP SIGNATURE-----
Version: 2.6.2

iQCVAwUBLtz1EjER5KvPRd0NAQEx7gP+IlVoJG1YVXKmQViVCtabX1owrH2MHDBg
MpKBq7T6NbPMTDUWLE7HNWTfw5BvZbSCC1uRRM2rKV6xHZPxU0buUsoDc5QLT10b
xYbs9/j81dlTve7/fMToJjNJuls61289XaOIlfPN+sBIGX1TwrtDKek6To8GsdAN
YmkUYUUFzL8=
=3fF9
-----END PGP SIGNATURE-----

In article <MAILQUEUE-101.941130155708.416@mhc.uiuc.edu>,
JEFF LICQUIA (CEI) <JLICQUIA@mhc.uiuc.edu> wrote:
>May I propose a "better" way (you be the judge here): Proxy the job.
...
>What say ye all?  I can tentatively volunteer my business account to do 
>the work (have to talk to my boss about it first, as that account has to 
>pay for volume and phone time).  I'll play with some code in the meantime 
>and see what I can come up with.

Now this is a good idea.   In order to successfully spoof a message, you
would have to block delivery of the spoofed message to the proxy checker
or block transmission of the proxy checker agent's warning.

If you're willing to write the code for it, I'm willing to provide the
machine on the Internet for it to run on.  It shouldn't take too much
effort, but I've already got a gazillion different pots in the fire as is
(not the least of which is getting some c'punk services up on hks.net).
-- 
Todd Masco     | "Roam home to a dome, Where Georgian and Gothic once stood
cactus@hks.net |  Now chemical bonds alone guard our blond(e)s,
cactus@bb.com  |  And even the plumbing looks good."  - B Fuller

From: "JEFF LICQUIA (CEI)" <JLICQUIA@mhc.uiuc.edu>

   Really, the only "unknown" with signed messages is whether they are valid 
   or not; it's pretty easy to distinguish the unsigned posts.  

The purpose of adding a header line to mark unsigned articles is _not_
to indicate that they aren't signed, it's to editorialize on the fact
that they're not signed.

There has been an argument that since marking doesn't accomplish
anything you couldn't already see, that it's useless.  Fine, the
premise is specious, because it's not intended to mark unsigned posts,
it's to comment on them.

   May I propose a "better" way (you be the judge here): Proxy the job.

A proxy should have it's own subscription list, which makes it an
opt-in system.  Other than that, I think a verifying proxy is a good
idea.

Eric