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

Re: SPAM vs. open list



On Sat, Feb 09, 2002 at 10:39:50PM -0500, Joey Hess wrote:
> > Frankly, its not such a terrible burden to sign up for a mailing list
> > and its not like -devel and -private are running low on mail from their
> > subscribers as it is.
> 
> So, put yourself in the shoes of an upstream author. You've written some
> program that is rather important and commonly used, and you get lots of
> bug reports and clueless whining and so on. And you know that debian
> includes your program, but that's really all you know about debian (you
> prefer a custom bsd-ports-like system you have maintained yourself for
> the past eight years, or slackware, or (god preserve us), red hat).  You
> have mixed (and perhaps ill-informed) opinions about debian for various
> reasons, and you've never really had much contact with the developer who
> maintains your package -- the few bug reports he forwarded to you have
> been utterly lost in the noise and you knew about the issues beforehand
> anyway.
> 
> Then you find out that we're doing something horribly wrong with your
> software. We have to be told how to fix this. So you try to get in touch
> with us by the first place you find, which is the debian-devel list
> mentioned prominently on the web site. And you get back some bounce that
> says, "Frankly, its not such a terrible burden to sign up for a mailing
> list (oh by the way, it only gets 200 messages a day or so)".
> 
> Well, what would you do?

In that case, I think I would be annoyed.  However, if the message was not
a bounce, but rather an informational reply telling me that my message was
being held for human approval because I was not a member of the list, and
was not a known poster, I'd probably understand that this sort of thing
happens - after all, my own lists have this exact setup.  The ones which
are annoying are the ones which force you to first subscribe, then resend
your message.

The thing is that a known poster list must be maintained (or people who
post from unsubscribed addresses all the time will have their mail held,
which is of course just annoying to them and the moderators both) and
people who haven't posted to a list for some reasonably long time should
be automatically removed from the list to prevent linear growth as time
passes.


Every time someone else suggests human moderators, the idea is shot down
because of the clear flaw regarding frequent non-subscribed posters.  I've
addressed that here.  This is not the first time I have suggested this
potential solution even since the new year, yet days after the last time I
suggested it on this list, people were still replying that human
moderators for non-subscriber mail would trip up anyone who doesn't use
their subscribed address for posts.

The meat of the automation already exists, unless since I stopped paying
attention to the ultruh sekrit content of #debian-devel Culus has stopped
tracking when developers were last seen doing _something_, be it uploading
a package, posting to a list, activity in the BTS, etc.  If all of that
can be tracked for every developer in the project, a few email addresses
can also be tracked with datestamps for automated housekeeping.


So the real question is, do we want to bellyache about spam on the Debian
lists, or do we want to do something productive about it?  I have spam
filters that catch nearly all the spam I get myself, but I choose not to
apply them to Debian lists because I only check the collection of messages
caught by it every three months or so.  It's not that big a deal to me,
given that most of the spam I recall was posted to security@, and I chose
not to resubscribe to debian-private, so I don't see it.  I'm a bit
annoyed that this comes up every couple of spams and have yet to see one
reason why my solution would not solve the problem and kill this thread
once and for all.  =p

-- 
Joseph Carter <knghtbrd@bluecherry.net>          If this sig were funny...
 
<Knghtbrd> hardcopy is for wussies
<Topher> computer program listings....next, on HardCopy

Attachment: pgpASQtjrNNTL.pgp
Description: PGP signature


Reply to: