gnunet-developers
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[GNUnet-developers] Fwd: amortizable hashcash paper


From: Christian Grothoff
Subject: [GNUnet-developers] Fwd: amortizable hashcash paper
Date: Sun, 27 Apr 2003 00:21:29 -0500
User-agent: KMail/1.4.3

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Hi guys!

I've one little extension that I would propose to the current 
directory/subspace proposal (http://www.ovmj.org/GNUnet/namespace.php3): 
popularity metrics. The idea is to use (amortized, averaged) hash-cash as a 
popularity metric for every signed file. It would work like this:
 
* user searches, findes signed (!) search result (SBlock)
* user downloads file, evaluates
* user can click on "vote" which will spend a bit of CPU time locally to
   vote -- if he liked the content
* each vote is associated with a voting-identity, and each user can have
   multiple of these; the vote is a hash code that has some number of
   bits that match the hash of the first n bytes of the search result (SBlock)
* the best k votes are stored at the end of the SBlock in the non-signed,
   non-encrypted portion of the block. They are propagated with the SBlock
   and merged with other votes, keeping the votes that had the highest
   number of colliding bits (see the paper below, this can be used to have
   an 'accumulated' vote that summarizes the overall popularity of the
   content).
* GNUnet keeps track of for which content a user voted, which OTHER user
  identities voted for similar content and weighs votes of other users by this
  "similar voting" behavior
* GNUnet can then *sort* search results by the overall score of the weighted
  votes

Users would have some incentive to vote since it will allow themselves to form 
communities of similar interest (the rock-music camp, the punk-music camp, 
the classic-music camp) and this will make spamming keywords much harder 
(nobody / few votes for the kitten collection). 

This will of course be an additional challenge for the implementation; for the 
first round I'd definitely just reserve an appropriate number of bytes at the 
end of the SBlock datastructure. 

I've attached a message from Adam Back in response to a discussion that Adam 
and myself had at the PET workshop about this.

Comments?

Christian

- ----------  Forwarded Message  ----------

Subject: amortizable hashcash paper
Date: Fri, 28 Mar 2003 08:02:52 +0000
From: Adam Back <address@hidden>
To: address@hidden
Cc: Adam Back <address@hidden>

this is the paper I mentioned which has applications for light weight
security distributed popularity metrics.

Aug 02 - "Hashcash - Amortizable Publicly Auditable Cost-Functions",
Tech Report, Adam Back

http://cypherspace.org/adam/hashcash/amortizable.pdf

Adam

- -------------------------------------------------------
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE+q2jZ9tNtMeXQLkIRAoFMAJ9S5s3FOPzyhgmUDDdDlaA7M7pE8wCglZBp
grlDVfoejt0R0GIU7BNbg7Q=
=/939
-----END PGP SIGNATURE-----





reply via email to

[Prev in Thread] Current Thread [Next in Thread]