On Sat, 4 Sep 2004, Sam Varshavchik wrote:
Hallam-Baker, Phillip writes:
That's fine. You might want to study the evolution of NNTP.
There, lack of
the official IETF stamp of approval did not inhibit NNTP from
evolving. If
you try to implement NNTP based only on IETF's standards-track
documentation, you won't get very far.
Are you saying that NNTP is a good example of protocol design?
Or a good example of creating a widely supported feature set?
NNTP works. It has survived the onslaught of spam far better than SMTP. I
may be see one or two spam message a week across all the newsgroups I'm
subscribed to. And NNTP's evolution was outside of the IETF process.
We haven't carried Usenet news for years, now. I'll have to check
somewhere see how spam-free Usenet is. And maybe its because many of the
groups are now moderated. Or maybe its because NNTP is particularly
efficient about spam and the cancel-bot** wars have stopped. All a
spammer needs to do is post one message a week to a group to make sure
that everyone in the group gets it once a week. Cross-posting the same
message to many groups means that you see it only once for all of those
groups.
Moderated mailing lists have also had less problem that unmoderated lists,
even though it is trivial to forge an on-list from address to post on a
(subscriber-only) moderated list. I've always found it strange that
abusers are so 'ethical' that they don't forge from addresses to get on
moderated lists.
--Dean
[** The cancel-bot wars were when usenet news admins would cancel spam
posts. The spammers would then repost the message. This process was
automated on both sides, and the result was thousands of spam messages to
a group. This made many groups unreadable. Spammers were blamed, but the
fault was just as much on the people who were doing the canceling, since
the groups were unmoderated and Usenet never had any kind of
non-commercial nature.]