[Cryptography] Proposal of a fair contract signing protocol

[Benjamin Kreuter]

On Mon, 2016-06-13 at 17:43 -0400, Allen wrote:
> > 
> > At any moment, both correspondents have a complete record of
> > everything
> > the other has acknowledged "on the record"
> Alice signs the contract and sends it to Bob for his signature.  A
> few
> months later, Bob claims he signed the contract and returned it the
> same
> day.  How do you propose proving the time that Bob signed the
> contract, if
> his email client can simply forge the time sent header, and Alice's
> can
> simply forge the time received, or forge the unacknowledged messages
> saying
> "Hey, where's the signed contract?"

There are various ways this is resolved in the offline pen-and-paper
case; e.g.:

* Introduce certification authorities (notaries) who assert that a
contract was signed no later than some time.

* Sign multiple copies and deposit some copies with other parties, who
can produce them on demand.

* Use certified mail or some other reliable delivery channel.

As someone pointed out earlier, the two-generals problem is inherently
hard to deal with.  People have had some awareness of this problem for
a very long time, long before the concerns of the Internet age, and
various solutions have been devised.  Treaties are signed in the
presence of witnesses and copies are sent to various countries.  The
USPS has long offered certified delivery and in some cases it is
required by law.

So for email, you would send a commitment to the signed contract to a
public mailing list or Usenet group, and if there is a dispute later
on, you can open the commitment.  The timestamps of list/Usenet
archives would prove that it was signed on or before some particular
date, and the binding property of the commitment would prevent
cheating.

-- Ben
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 819 bytes
Desc: This is a digitally signed message part
URL: <http://www.metzdowd.com/pipermail/cryptography/attachments/20160614/98dd95b5/attachment.sig>