[Mailman-Users] Reply to list

[Brad Knowles]

At 2:43 PM -0800 2004/02/05, Mark Dadgar wrote:

>  Yes, but in the text to which I was referring (that you snipped in
>  your reply above), you stated that it's easy enough to solve Reply
>  All problem of ending up with the original sender and the list
>  address in the To: line by just deleting the original sender's
>  address.  But now you're telling me that's not an option in a
>  large number (in terms of user base) of mail clients.
>
>  So, pick one.  Which way do you want it?

	In the clients I've seen, the original sender would get put in 
the "To:" field, and in a reply that would be unchangeable.  All the 
other addresses would get put in the "Cc:" field, and that would be 
changeable.

	Setting "Reply-To:" to the list would cause the reply to be sent 
only to the list, and that would be in the unchangeable "To:" header. 
Not setting this field at all would allow the reply to go back to the 
original poster, with the other addresses changeable.

>  That's pretty cool.  BUT, how many of those 5 million plus lowest
>  common denominator people did you actually interact with?

	Quite a lot.

>  But in any case, this is irrelevant.  I don't think you can
>  dictate behaviour to everyone based on your sampling.

	No, but I can tell you that I've seen behaviour on many, many 
occasions which you don't ever appear to have seen, and in my 
experience that is typically explained by the size of the user base 
in question.

>  How often does that really happen (that sensitive information is
>  shared).  I mean, really.

	Quite frequently, actually.  Moreover, just once is more than enough.

>  It could.  Nice corner case.  And I would not advocate turning Reply
>  To List on in a circumstance like that.

	It's actually a quite common case, not a corner.

>  But in the rest of the majority of cases, where accidentally
>  replying to the list broadcasts nothing more seriously than
>  your Aunt Martha's secret carrotcake recipe, the risk is worth
>  living with in exchange for the myriad benefits.

	What "myriad benefits"?  I haven't heard a single one.

>  Don't you think this is a bit of a stretch?  We're not talking
>  about air traffic control here.

	No, not really.  You really can lose your job over these kinds of 
things.  Someone losing their life is not out of the question.

	In essence, you really are playing a form of Russian Roulette.

>  At the end of the day, though, it doesn't really matter what
>  the size of the list is, PROVIDED that you can tailer the
>  software appropriately to your intended use.  You are
>  advocating making that impossible.

	No, I'm not.  I am not advocating that this feature be removed 
from Mailman.  I am advocating that this feature be left turned off, 
unless an experienced administrator (who fully understands the issues 
involved) decides that it is necessary for their particular purpose. 
I can't imagine what that is, and I am kind of hesitant to give 
people tools to shoot their feet off casually, but I do recognize 
that there might be situations where it could be useful.

	However, from what I've seen from you so far, you do not have 
enough experience to fully understand all the issues.


	Maybe it should be a feature that requires changing the source code....

>  I love arguments like that.  Nonsensical, but completely irrefutable.

	Obviously, you haven't lost a job over situations like this. 
Until something of that scale happens to you, you may well not 
understand.

-- 
Brad Knowles, <brad.knowles at skynet.be>

"They that can give up essential liberty to obtain a little temporary
safety deserve neither liberty nor safety."
     -Benjamin Franklin, Historical Review of Pennsylvania.

GCS/IT d+(-) s:+(++)>: a C++(+++)$ UMBSHI++++$ P+>++ L+ !E-(---) W+++(--) N+
!w--- O- M++ V PS++(+++) PE- Y+(++) PGP>+++ t+(+++) 5++(+++) X++(+++) R+(+++)
tv+(+++) b+(++++) DI+(++++) D+(++) G+(++++) e++>++++ h--- r---(+++)* z(+++)