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

Re: [Asrg] 6. Proposals - Pull System (revisited)



On 2003-11-26 09:36:34 +1030, Chris wrote:
> >
> > It is worth noting that the word spam was coined in response to a failure
> > that appeared in NNTP, a pull system long before the problem spread to
> > email.
> >
> 
> 
> NNTP does allow a news server to ignore mail originating from known
> spam/objectionable sources
> 
> it is the CHOICE of news server administrator to accept or deny from a
> source
> 
> many news servers refuse to accept certain newsgroups
> 
> this is an example of a *pull* system functioning perfectly
> 
> and a good example of why pull should be a serious contender for use in
> e-mail

I disagree. NNTP is not generally a pull system. News between servers
are mostly transported with a push system, only clients (and small
servers like leafnode, hamster, etc.) use it as a pull system. This is
very similar to E-Mail: Mails are transported to the user's mail server
with a push system (SMTP) and fetched from there with a pull system
(POP, IMAP). NNTP as a protocol just supports both modes.

There are some differences between SMTP and NNTP feeds:

SMTP is a unicast/multicast (oligocast?) protocol, NNTP is broadcast
protocol. You only have to send a single NNTP message to reach all
readers of a newsgroup, but you have to send a mail message to every
single email address if you want to reach many people. So a newsgroup
spammer sends out at most a few thousand messages (one for each
newsgroup), but a mail spammer sends out millions. NNTP traffic is also
generally public, so it is easy to set up bots which monitor the traffic
and send out cancels, or to set up content filters. I think this is why
Usenet spam died out after some time (at least I see very little Usenet
spam these days - maybe I just read the wrong newsgroups :-).

> > The issue is not how the bits are shipped. Every email communication is by
> > its nature initiated by the sender. It is the ability to initiate without
> > any form of authorization that creates the spam problem. In effect the
> > sender can consume unbounded respources. I don't see how tweaking the
> > details of the protocol affects this.
> 
> you miss the point
> 
> if the recipient refuses the email it does not even get onto the transport
> system
> 
> just a small portion of it

Which reduces the traffic but not the time users spend. If I only see
the sender address and the subject, I may still have to look at the
content of the message to determine whether it is spam. There's not much
difference whether it is in my local mailbox or on another server
(except that accessing a remote server is slower). OTOH, if my mail
filter has the whole message, it has a very good chance to determine
whether it is spam or not (So I won't have to look at it any more).

	hp

-- 
   _  | Peter J. Holzer    | In this vale
|_|_) | Sysadmin WSR       | Of toil and sin
| |   | hjp@hjp.at         | Your head grows bald
__/   | http://www.hjp.at/ | But not your chin.           -- Burma Shave

Attachment: pgp00100.pgp
Description: PGP signature