The Privacy/Untraceability Sweet Spot

[Eric Murray]

On Fri, Aug 31, 2001 at 07:25:27PM -0700, georgemw at speakeasy.net wrote:
> On 31 Aug 2001, at 17:07, Eric Murray wrote:
> 
> > The various CDRs should deal with this-- they keep track of
> > which Message-IDs they have posted, and don't post messages
> > that they have already posted.  They do pass those messages
> > on to the other CDRs that they peer with though, so by posting
> > two messages you're adding to the number of messages
> > that get sent on the "backbone"...
> > 
> 
> Is it possible to set things up so that duplicate messages are 
> filtered out based on a message digest rather than a message ID?

Just about anything is possible.

> I'm dumb as a post and have no clue how message IDs are 
> generated in the first place,  but it seems to this simpleton that
> any hash function on the message body would be guarenteed to 
> catch duplicate messages.

Message-ID is generated by the originating MTA (that's the ISP's
sendmail or whatever program they're using to send mail).

It's supposed to be unique-- it's often something like

200109012124.f81LObL18955 at slack.lne.com

where there's a date component (200109012124) and the sending
machine's name, so that it's unique in a global namespace.
Of course there's nothing keeping someone from sending
mail with whatever message-id they want, assuming that they
can control their sending MTA.

The reason that it's used for identifying already posted messages
is that procmail, which the CDR system is based on, has a nice
built-in hook for keeping a message-id database and identifying
duplicates.


Eric