On Tue, 12 Oct 1999, Steve Bergeon wrote:
> But the lists are public, and meant to be joined in an automated
> manner. Again I am asking what acceptable limits may be, one
> person/list/server? Two? Three? Is it considered some sort of spam
> (obviously not UCE) to go over this threshold?
I have to sort of agree with Steve here. Mailing lists systems are
automatic systems designed to automate or semi-automate subscription
activity (as well as other things).
But there is still a massive assymetry. Many lists are set up to involve
some human intervention by the list manager for subscription (to approve
subscriptions). So a scripted subscription attempt that hit say, 10000,
lists might hit 1000 closed subscription lists, and in the best of
circumstances (no suspicion of attack) would require separate manual
action by 1000 different list managers.
> Were you in fact called out of bed because of the number of mails,
> or was it notification of a loss of system resources related to
> the mail? Does your pager go off when somebody joins a list? Two?
There certainly was no problem with the load (and if there was that would
be my problem. Our system should easily carry a bunch of simultaneous
subscription attempts). I was not woken up by this, but I certainly cost
me a real amount of time once I did notice it. It's not the number of
subscription requests, but the indiscriment nature of them. I saw an
address subscriping and attempting to subscribe to lists at Cranfield
indiscriminantly. At the time, I could imagine no non-malign reason why
anyone would do that. It's not the number of lists, it is the manner in
which they are selected.
Jeffrey Goldberg +44 (0)1234 750 111 x 2826
Cranfield Computer Centre FAX 751 814
Relativism is the triumph of authority over truth, convention over justice.