// COMPLETE THREAD

Mainstreaming PGP on Usenet

9 expanded posts ยท every known parent and child

NODE 4278ffc0Mainstreaming PGP on Usenet
I'm surprised nobody has brought this up before....

                     FIRST CALL FOR VOTES (of 2)
            moderated group soc.culture.russian.moderated

<snip>

6. Individuals in the database of known readers may post freely to
the group, subject to the conditions in sections 1-4 above.  If need
arises, the robomoderator may perform PGP verification of the
identity of the known reader and, if the reader requests so,
automatically reject all the submissions from the reader without
a valid PGP signature.

--
Bruce Baugh
bruce@aracnet.com
http://www.aracnet.com/~bruce
NODE f42b5058Re: Mainstreaming PGP on Usenet
comp.os.ms-windows.announce will also start using PGPMoose when I get
around to it, probably today. 

See http://www.sydney.sterling.com:8080/~ggr/pgpmoose.html

-rich
 win-request@metrics.com

On Sun, 3 Mar 1996, Bruce Baugh wrote:

> I'm surprised nobody has brought this up before....
> 
>                      FIRST CALL FOR VOTES (of 2)
>             moderated group soc.culture.russian.moderated
> 
> <snip>
> 
> 6. Individuals in the database of known readers may post freely to
> the group, subject to the conditions in sections 1-4 above.  If need
> arises, the robomoderator may perform PGP verification of the
> identity of the known reader and, if the reader requests so,
> automatically reject all the submissions from the reader without
> a valid PGP signature.
> 
> --
> Bruce Baugh
> bruce@aracnet.com
> http://www.aracnet.com/~bruce
NODE 4f6fadbdRe: Mainstreaming PGP on Usenet
Just Rich <rich@c2.org> writes:
> comp.os.ms-windows.announce will also start using PGPMoose when I get
> around to it, probably today.

All moderated newsgroups should use it.

However the robomoderator that Igor Chudov wrote uses digital signatures to
authenticate _posters as follows: there's a "white list" (as opposed to "black
list") of trusted posters whose submissions will be approved and posted
automatically, without going through any moderator. A person on the "white
list" can request that the robomod check his signature and not post it if the
submission may be a forged e-mail. I strongly urged Igor to make digital
signatures mandatory for "white list" membership, but he argued that a lot of
likely posters lack the brains to use PGP. The robomod will look at things like
"Received" lines in the e-mailed submission to try and detect forgeries.

While at it, here's a pre-filled ballot. Just add your name and e-mail it
to russian-vote@netagw.com.

===== BEGINNING OF BALLOT: Delete everything BEFORE this line =====
-------------------------------------------------------------------
soc.culture.russian.moderated Ballot
<SCRM-0001> (Do not remove this marker!)
-------------------------------------------------------------------
Please provide your real name, or your vote may be rejected.  Place
ONLY your name (ie. do NOT include your e-mail address or any other
information; ONLY your name) on the next line beside "Voter name:".
Voter name:
-------------------------------------------------------------------
Insert YES, NO, ABSTAIN, or CANCEL inside the brackets for each
newsgroup listed below (do not delete the newsgroup name):

 Your Vote   Newsgroup
 ---------   ---------
[ YES     ]  soc.culture.russian.moderated (moderated)
-------------------------------------------------------------------
===== END OF BALLOT: Delete everything AFTER this line ============

Vote YES to see how a robomoderator that uses PGP works out.

---

Dr. Dimitri Vulis
Brighton Beach Boardwalk BBS, Forest Hills, N.Y.: +1-718-261-2013, 14.4Kbps
NODE 4d87a9aaRe: Mainstreaming PGP on Usenet
-----BEGIN PGP SIGNED MESSAGE-----

To: cypherpunks@toad.com
Subject: Re: Mainstreaming PGP on Usenet

Dr. Dimitri Vulis wrote:
> 
> Just Rich <rich@c2.org> writes:
> > comp.os.ms-windows.announce will also start using PGPMoose when I get
> > around to it, probably today.
> 
> All moderated newsgroups should use it.
> 
> However the robomoderator that Igor Chudov wrote uses digital signatures to
> authenticate _posters as follows: there's a "white list" (as opposed to "black
> list") of trusted posters whose submissions will be approved and posted
> automatically, without going through any moderator. A person on the "white
> list" can request that the robomod check his signature and not post it if the
> submission may be a forged e-mail. I strongly urged Igor to make digital
> signatures mandatory for "white list" membership, but he argued that a lot of
> likely posters lack the brains to use PGP. The robomod will look at things like
> "Received" lines in the e-mailed submission to try and detect forgeries.

This is not exactly correct. The check for white list will be there
exactly as you specified. It will have a toggle switch, so that we
can turn it on and off, depending on how badly we are hit with
forgeries.

> 
> While at it, here's a pre-filled ballot. Just add your name and e-mail it
> to russian-vote@netagw.com.
> 

Please do NOT use this pre-filled ballot to vote (although of course
I would be pleased with cypherpunks helping to let such an experiment
go). According to voting rules, you have to vote only in response to the
full CFV posted to news.groups. If you do not have access to news or
your news system expired the CFV already, you can retrieve the CFV
(containing the rationale, charter and the ballot) by sending email to
russian-cfv-request@netagw.com. Body of your message will be ignored.

All votes that use pre-filled ballots are invalid.

	- Igor.

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

iQCVAwUBMTqH5cJFmFyXKPzRAQGYVQQAktT4DYcIcDCV9CSWK+BYGPGS9S609m59
whde9eCtG7d6XRUdnAlZnJQ/sqr/TXVtpfTfGYTZJnD0HYOO0INq7+jN7qHs/7ue
KKQMAHM0mi9njEsKUP1cFvn+h68UNDSlH9zrjuMgLZvPxHcal+Wg0gAT9MBsO3xs
4HG2DcpBTSo=
=JUnd
- -----END PGP SIGNATURE-----

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

iQCVAwUBMTqIK8JFmFyXKPzRAQFRFwQAk7/feGthnWXKKJIH8m4XTRbfW0wT1dmI
GazD0DSLU2Yy/L31QeIA2wdXqxHRIv15SXqv4/rdRnpbbRnxEmIO3jzZmRjvjWTK
+wXO/kDrlmiiO+QCmg6jQs8BD4Mk4wNkqwsgUhxLnX9z6nwRA3KUqpOMp8Y45HRT
aFyQV5SYByg=
=cTQ0
-----END PGP SIGNATURE-----
NODE d4b2cf50Re: Mainstreaming PGP on Usenet
Bruce Baugh <bruce@aracnet.com> writes:

> I'm surprised nobody has brought this up before....
>
>                      FIRST CALL FOR VOTES (of 2)
>             moderated group soc.culture.russian.moderated
>
> <snip>
>
> 6. Individuals in the database of known readers may post freely to
> the group, subject to the conditions in sections 1-4 above.  If need
> arises, the robomoderator may perform PGP verification of the
> identity of the known reader and, if the reader requests so,
> automatically reject all the submissions from the reader without
> a valid PGP signature.

Igor Chudov, who coded the robomoderator, reads Cypherpunks, and
is known to appreciate and use good suggestions.

One issue that hasn't been addressed by the s.c.r.m robomod is the
possibility of persistent nyms: that is, Alice D. Nonymous somehow makes
her public key known to the robomod; and later if someone submits an article
via some anon remailer claiming to be hers, it would be rejected if the
signature doesn't check. Of course, her true submissions would be accepted
from any remailer. How could such protocol be implemented?

(Of course, some people have what they believe to be valid reasons not
to use PGP.)

---

Dr. Dimitri Vulis
Brighton Beach Boardwalk BBS, Forest Hills, N.Y.: +1-718-261-2013, 14.4Kbps
NODE 030d42c0Re: Mainstreaming PGP on Usenet
-----BEGIN PGP SIGNED MESSAGE-----

To: dlv@bwalk.dm.com (Dr. Dimitri Vulis)
Cc: cypherpunks@toad.com
Subject: Re: Mainstreaming PGP on Usenet

Dr. Dimitri Vulis wrote:
> Bruce Baugh <bruce@aracnet.com> writes:
> > I'm surprised nobody has brought this up before....
> >                      FIRST CALL FOR VOTES (of 2)
> >             moderated group soc.culture.russian.moderated
> >
> > <snip>
> >
> > 6. Individuals in the database of known readers may post freely to
> > the group, subject to the conditions in sections 1-4 above.  If need
> > arises, the robomoderator may perform PGP verification of the
> > identity of the known reader and, if the reader requests so,
> > automatically reject all the submissions from the reader without
> > a valid PGP signature.
> 
> Igor Chudov, who coded the robomoderator, reads Cypherpunks, and
> is known to appreciate and use good suggestions.

Indeed. If you indicate your interest, I can post here a more or less
full description of the robomoderator, how it implements secure
exchange between itself and human moderators, verifies submissions,
and signs approved articles for posting (it uses PMApp by Greg Rose).

Your criticisms will be most welcome.

> One issue that hasn't been addressed by the s.c.r.m robomod is the
> possibility of persistent nyms: that is, Alice D. Nonymous somehow makes
> her public key known to the robomod; and later if someone submits an article
> via some anon remailer claiming to be hers, it would be rejected if the
> signature doesn't check. Of course, her true submissions would be accepted
> from any remailer. How could such protocol be implemented?

We see no problem with user posting under pseudonyms, as long as they do
not try to pretend to be other real people and do not constantly mutate,
and submit their messages to the robomoderator for consideration.

> (Of course, some people have what they believe to be valid reasons not
> to use PGP.)

Sorry if this question will provoke a mini flamewar, but what are
such reasons?

	- Igor.

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

iQCVAwUBMTp458JFmFyXKPzRAQEbQAP8CAtCiNm9h7pijz4+qrm6FFGRBDjsqvZb
RkgFetA01oEONdp+RW3NP6GBY1zeNg7+HahfoavNPhASwBl230hLni1fEW5pL75o
J5v5yNCYT4/N1aVfchoo50kjXv+KqvRyjr5YNxVmd/IFKokSV1w9ASfdXVk/7uDB
Ep1n1jmP4OQ=
=EQOd
-----END PGP SIGNATURE-----
NODE 4ec8a6b3Re: Mainstreaming PGP on Usenet
On Sun, 3 Mar 1996 ichudov@algebra.com wrote:

> Dr. Dimitri Vulis wrote:
...
> > Igor Chudov, who coded the robomoderator, reads Cypherpunks, and
> > is known to appreciate and use good suggestions.
> 
> Indeed. If you indicate your interest, I can post here a more or less
> full description of the robomoderator, how it implements secure
> exchange between itself and human moderators, verifies submissions,
> and signs approved articles for posting (it uses PMApp by Greg Rose).

It sounded cool (as heard on the moderator's list), but too complex for my
needs, and I think it required some stuff I don't have. Of course, for a
higher-traffic group, it's worth it. I don't see a way around the problem. 

... 
> > (Of course, some people have what they believe to be valid reasons not
> > to use PGP.)
> 
> Sorry if this question will provoke a mini flamewar, but what are
> such reasons?

1. If you're like me and you habitually read your mail online on a host on
the Internet, no matter how secure, then that's a security risk. Of course
one could, and many people do, create multiple PGP keys, one for casual
authentication and encryption online, and another held in check for stuff
that needs to be secure. I just don't bother with a "10% secure" key for 
cpunks and casual mail, though I do sign most of my Usenet posts.

2. If you send a PGP-signed message to a non-PGP-aware list, there will 
be questions, and sometimes ridicule. Sometimes this is an opportunity 
for education, sometimes it's just not worth the trouble.

3. Using PGP may attract the unwelcome attention of hostile local or
foreign governments, or possibly other armed thugs. 

4. "It's too hard."

-rich
NODE 2ff09ad8Re: Mainstreaming PGP on Usenet
ichudov@algebra.com (Igor Chudov @ home) writes:
> Dr. Dimitri Vulis wrote:
...
> > One issue that hasn't been addressed by the s.c.r.m robomod is the
> > possibility of persistent nyms: that is, Alice D. Nonymous somehow makes
> > her public key known to the robomod; and later if someone submits an articl
> > via some anon remailer claiming to be hers, it would be rejected if the
> > signature doesn't check. Of course, her true submissions would be accepted
> > from any remailer. How could such protocol be implemented?
>
> We see no problem with user posting under pseudonyms, as long as they do
> not try to pretend to be other real people and do not constantly mutate,
> and submit their messages to the robomoderator for consideration.

On the contrary, I was thinking of a situation when a person is posting through
an anonymous remailer, yet wishes to establish a persistent nym that can't be
impersonated. E.g., someone may submit articles via remailers (different every
time) and have a signature 'Alice D. N.'; what's to prevent someone else from
submitting an article and also signing it 'Alice D. N.'?

I was thinking of allowing the user to add a 'From: <nym>' in the first
paragraph of the PGP-signed block. To establish the nym, Alice would first
post her public key under the name of Alice; then she would post things like

From: remailer@somewhere

-- begin pgp signed msg

From: Alice

...

This would also address the problem of someone's misconfigured system where
his submissions appear to come from moron@camelot.ptu.edu or
moron@pendragon.ptu.edu or some other random hostname.

P.S. So, when are you setting up your own mixmaster remailer, Igor? :-)

---

Dr. Dimitri Vulis
Brighton Beach Boardwalk BBS, Forest Hills, N.Y.: +1-718-261-2013, 14.4Kbps
NODE 93f76908Re: Mainstreaming PGP on Usenet
-----BEGIN PGP SIGNED MESSAGE-----

To: dlv@bwalk.dm.com (Dr. Dimitri Vulis)
Cc: cypherpunks@toad.com
Subject: Re: Mainstreaming PGP on Usenet

Dr. Dimitri Vulis wrote:
> 
> ichudov@algebra.com (Igor Chudov @ home) writes:
> > Dr. Dimitri Vulis wrote:
> > > One issue that hasn't been addressed by the s.c.r.m robomod is the
> > > possibility of persistent nyms: that is, Alice D. Nonymous somehow makes
> > > her public key known to the robomod; and later if someone submits an articl
> > > via some anon remailer claiming to be hers, it would be rejected if the
> > > signature doesn't check. Of course, her true submissions would be accepted
> > > from any remailer. How could such protocol be implemented?
> >
> > We see no problem with user posting under pseudonyms, as long as they do
> > not try to pretend to be other real people and do not constantly mutate,
> > and submit their messages to the robomoderator for consideration.
> 
> On the contrary, I was thinking of a situation when a person is posting through
> an anonymous remailer, yet wishes to establish a persistent nym that can't be
> impersonated. E.g., someone may submit articles via remailers (different every
> time) and have a signature 'Alice D. N.'; what's to prevent someone else from
> submitting an article and also signing it 'Alice D. N.'?
> 
> I was thinking of allowing the user to add a 'From: <nym>' in the first
> paragraph of the PGP-signed block. To establish the nym, Alice would first
> post her public key under the name of Alice; then she would post things like
> 
> From: remailer@somewhere
> 
> -- begin pgp signed msg
> 
> From: Alice
> 
> ...
> 
> This would also address the problem of someone's misconfigured system where
> his submissions appear to come from moron@camelot.ptu.edu or
> moron@pendragon.ptu.edu or some other random hostname.
> 

As far as I understand, the problem boils down to this:

   Nym users want to have an identity (belief of other users that
   a set of articles originating from many anonymous addresses were in
   fact written by one person). At the same time, we want to prevent
   users who do not have a permanent return address from using
   addresses of other persons.

One of the problems with allowing users to specify return addresses in
the letters is a possibility of forgery: what if
Mallory@evil-services.com posts a MAKE MONEY FAST message, and specifies
that her "From: " address should be Popugaev@get_high.edu? Such posting
could get Mr. Popugaev in trouble.

Maybe the following rewriting rule may be a good compromise between
functionality and security, for PGP signed messages:

1. Original "From: " address is rewritten as "X-Origin" or some such.

2. "From:" address is always set to the main user ID of the PGP key that
was in the signature.

3. For "Reply-To: " we use "Reply-To: ", if it is present, then we try
"From: ", and if "From: " is not present, "Reply-To: " is not set.

4. If "Subject: ", "Date: ", "Message-ID: ", or "Newsgroups: " is
present on the block of pseudo-headers starting with the first line of
the text, use them instead of trusting the headers of an email.

This way, we achieve the following results: 

1. Positive and reliable identification of users is possible to every
reader of soc.culture.russian.moderated, not only to moderators;

2. Users gain additional protection from man-in-the-middle attacks by
using well protected pseudo-headers withing PGP signed blocks; they
cannot misuse this feature by lying about who they are;

3. Those who do not want to bother do not have to;

4. People with misconfigured email addresses may have at least
some address field ("From: ") set correctly.

5. Nyms can post freely through any anon remailers and always have their
identity show up in the "From: " field, even if remailers do not allow
users to specify their identity at all.

Note that I agree that we need to have a database of MD5 checksums of all
submissions and carefully process duplicates.

What do you think?

	- Igor.

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

iQCVAwUBMTvi6MJFmFyXKPzRAQGEMQP8C4V9gCs5REc5hez0gRP7bXn9NGV5S/6l
fxJo4SPmCBdWxn+msLxchbrho/hlhcUMaPuswcnacgrqEAyd1H4yIiMyZ1s6z06e
0q6WQ8QUy/E1nrc4lCSXKUBYB8MV/SGlynxxq3X9T2eF2lmnoArWj4QpfcVgk9RR
HvcvpK3GWuA=
=OXCv
-----END PGP SIGNATURE-----