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

META: mailing-list technicalities (Was Re: All of uu.net about to be banned.)



In debian-devel andreas wrote:
>>
>>  'Deny all incoming, execpt those that are subscribed'
>
>germany, karlsruhe has a ka.* local news hirarchy with ka.lists.linux*
>containing many mailing lists of vger.rutger.edu (we got an ok), debian
>and other stuff like bugtraq. we many people here are not subscribed,
>but use news to read, but email to reply/follow up.

So do I. Large-volume mailing lists need handling features akin
to that of news. Setting it up, of course, can be delicate.  I
*once* disregarded security procedures (server had been running
perfectly for months, and I just wanted to make one itty-bitty
change to the script) and immediately managed to feed (*one*)
mail back to a list. Furrfu. Still smarts. Still haven't figured
out why itty-bitty change had unwanted side effect. Changed
everything to use heavy-duty human checking so that it can not
happen again.

>that setup is very nice, and i guess many people are useing it.

These are old problems, and the solutions have been thrashed out
many times already.

The AFAIK normal, accepted solution to implement a
wide-readership abuse-proofed list is

	1) confirm subscriptions by auto-bot
	2) restrict posters to:
		a) addresses subscribed to list
		b) addresses subscribed to list-posters

list-posters does not receive any mail, it is only for people
who write using different mail addresses than they use to
receive list mail. Even if I didn't use inn to receive my
mailing-lists I would do that.

I suppose that you wouldn't really need confirmation for poster
subscription, but why not.

Another (complementary) option would be to send posts from
non-subscribed addresses to a moderator, but that has the
obvious disadvantage of needing a moderator, and would not
encourage people to subscribe to the poster list.

-- 
#include <std_disclaim.h>                           Lorens Kockum


Reply to: