[hashcash] Re: stamp creation std. deviation

  • From: "Eric S. Johansson" <esj@xxxxxxxxxx>
  • To: hashcash@xxxxxxxxxxxxx
  • Date: Fri, 13 Aug 2004 09:56:59 -0400

Jean-Luc Cooke wrote:

This policy I'd personally like to see it as "hashcash recomands X bits, but
it is up to the mail-server/procmail user to decide how to filter mail".

Reasons:
 1) Mailing lists like linux-kernel may find ti too expensive to compute
 hashcash values for 3,000 emails 400 times a day.
 2) lots of others

I wish people would get off of the mailing list issue. It's not even a dead horse for people to beat anymore, its fragments of skeleton's.


It's generally understood that mailing lists can't afford to generate stamps. No problem. The camram project has proposed but not been able to experiment with an alternative stamp/peer to peer identification system using asymmetrical key systems.

As for leading servers decide how to filter etc. etc. for difference postage rates, I think that's potentially a very bad idea. how do you propose to handle postage deltas? I'm sending out a postage due notice and putting responsibility back on the sender. If you plan on doing the same thing, we should have an agreed-upon standard mechanism for communicating what postage is required as well as that the postage due notice is in response to a real message. We should also have a mechanism for allowing non hashcash enabled users to generate postage.

My fundamental philosophy in all of this is to minimize what the user sees because spam filtering is not their goal in life. It's something they do because they have to like filling the gas tank in the car or checking the oil. These are not fun things. These are not activities people live for. They're just dumb things you gotta do. therefore, you make them as innocuous and as minimal as you possibly can.

so let's figure out ways to automate this process as much as possible and win on the anti-spam front not just from higher effectiveness but from higher usability. If you don't believe you can make an interest and interface drop dead easy-to-use, take a look at the camram interfaces.

Also, 25bits is very high IMHO.

not really. We are looking for 15 to 30 seconds stamps to start. This puts a serious cramp in the profit model of a spammer. If you adopt the camram model of stamp generation (i.e. stamps required when communicating with people you don't know) then the average load drop significantly and higher bit value stamps are much more practical.


---eric


-- Speech recognition in use. It makes mistakes, I correct most

Other related posts: