> I am interested in a similar thing.  I would like to have a system which
> will allow public keys to verified against several different criteria.
>
> 1) That a public key was verified by an actual human an not just created
> automatically.  By this I mean that after a public key is generated an
> actual human went though a fairly involved process to verify it.  This
will
> involve a variety of tasks which will be very difficult for a computer to
> automate.  Once the human complete the process the verifier(s) will sign
> the public key.  The primary motivation behind this is to allow anyone to
> post messages to a group but prevent spam by disallow certain public keys
> of known spammers.  This type of list will be community maintained by some
> sort of voting system.  The voters must also have verified keys and may
> vote on other voters to prevent abusive voters.

Though I can see the usefulness of such a system in building trust etc in
online communities, I can't (off the top of my head) think of any way you
can require some form of manual verification - if the function can be
performed by humans, it's going to be difficult to check the validity of it
by computer, and difficult to require it for a key.

> 2) Being able to securely attach email address and other forms of
> out-of-band identification information to a public key.  Obviously the
> reason for doing this is to make a anonymous public-key not so anonymous.
> This should be easier to do since the only real challenge is doing so
where
> no one particular node can be trusted.  Maintaining anonymity in this case
> is obviously not important.

This would be an integral part of the network I was envisaging. The
particular structure best suited to trusted key-exchange may not acutally
require anonymity, in fact in some cases it might be better without it - for
example, tracking down a node that is 'poisoning' the net by looking for the
common node in all paths that result in 'poisoned' data.

Another problem that arises is that although a network such as I was
suggesting can ensure data entered at one end can be stored & transmitted
throughout the network with a low probability of modification (and a
statistic that indicates the chance of that) and a high probability of
discovery if a subset of nodes are modifying, there is no way the network
can determine the validity of the data that was entered in the first place -
for example, what do you do if you use the network to look up the pubkey of
someone you want to contact, and it turns up more than one key, at least one
of which could be the key of an intruder trying to listen to your
conversation?

Nick

> Do you have any ideas?


_______________________________________________
freenet-tech mailing list
[EMAIL PROTECTED]
http://lists.freenetproject.org/mailman/listinfo/tech

Reply via email to