Re: uol.com.br and petsupermarket
On Wed, 15 Mar 2006, Lars Wirzenius wrote:
> * Ask someone in Brazil to call up the petsupermarket people on the
> phone and see who there is interested in Debian and if they could please
> not forward their Debian list subscriptions to uol.com.br?
It has been attempted already, and it was not well received. Besides, it
wouldn't do much good even if petsupermarket was eager to help. It is UOL
which is the real source of the trouble, and unless the owner of
petsupermarket is friends with someone from the UOLs board, they will ignore
him just like they ignore everyone else anyway.
I'd advise against contacting petsupermarket again.
> * Do a big "spring/fall cleanup" and unsubscribe *everyone*, with due
> advance notice of course. The linux-kernel list did that at least once,
> many years ago. This is different from kicking off only uol.com.br
> subscribers, since it will get rid of the offending address (and also
> any other people who are too lazy to unsubscribe, of course). Admittedly
> it is highly radical. (When people re-subscribe, keep time stamps of
> when they do that, to correlate with if the problem still comes back.)
That would be a good thing, but not exactly because of the UOL backscatter.
We would get rid of most of the address-harversting bots for a while, plus
we'd get rid of a lot of cruft addresses that are still subscribed but lead
But we must make sure that resubscribing should be subject to a
non-braindead system where you can't just bounce the message back to be
subscribed. Automated systems do that when they are not smart enough to
know that you should not auto-reply to precedence: list or precedence: bulk.
Heck, a badly designed DSN could get a borked MTA subscribed if all it takes
is a reply with no changes...
"One disk to rule them all, One disk to find them. One disk to bring
them all and in the darkness grind them. In the Land of Redmond
where the shadows lie." -- The Silicon Valley Tarot