[Mailman-Users] Reply to list

Brad Knowles brad.knowles at skynet.be
Thu Feb 5 23:27:35 CET 2004

At 1:39 PM -0800 2004/02/05, Mark Dadgar wrote:

>  You just told me in a previous email that some people cannot edit
>  the To: line on a reply.  So which is it?

	On the software I've seen, you can't change the To: line in a 
reply, but you can add other addresses in the Cc: line.

>  Then we have at least equivalent experience.  Except that I appear
>  to live in the real world and you live in a techy vacuum.  There is
>  nothing wrong with that, mind you, as long as you realize that the
>  assumptions you make about people's usage patterns are colored by
>  the people you interact with.

	Well, I was the Sr. Internet Mail Systems Administrator for AOL, 
and the first Internet Mail Operations person they ever hired.  I've 
worked in environments supporting five million plus users, most of 
whom I believe everyone would agree are pretty 
lowest-common-denominator -- they don't get any lower.  I've also 
worked with the Listserv/LSMTP installation at AOL, and I helped set 
up many of the machines that were being used by the mailing list 
administration staff.

	How many millions of lowest-common-denominator users have you supported?

>  Chuq made the point that he had done exhaustive research on the way
>  people use this stuff.

	Indeed, he has.  He doesn't like to brag about it, but he does 
run some of the largest known Mailman mailing lists, and his systems 
are on the same scale as the Kolstad & Chalup papers that I have 
previously mentioned on this mailing list.  Chuq could probably write 
the third installment in "How to Manage Very Large Mailing Lists".

>                          The fallacy he has unknowningly fallen into,
>  and you have as well, is that the behaviour of the software needs
>  to be tailored to the user base.  There is no single case that solves
>  all problems.

	True enough.  But in this case, there is a particular 
configuration which is known (and proven) to reduce the amount of 
information being needlessly destroyed by the mailing list 
administrator, and which is likely to result in the least amount of 
embarrassment if someone just hits the reply key and then shares out 
all sorts of really sensitive information.

	Your method would have all that incredibly sensitive information 
automatically sent back to the entire mailing list, which could 
certainly cost someone their job.  That kind of behaviour could 
conceivably cost someone their life.

	If it was your job and your life that was on the line every 
single time that one of your customers hit the reply key, which would 
you want?  Do you really want to play Russian Roulette with 20,000 
other people?

>  However, you are advocating a reduction in choices by doing away
>  with Reply To List.  And in my experience it just is not as big
>  a problem as you people think.

	Then I would have to conclude that you haven't been doing this 
for long, or at least not with any mailing list of any real size.

>  The poster (Paul?) who said it's probably appropriate for 100
>  person lists but not for 20K person lists was right on the money.

	See above.  Russian Roulette with 100 people is more likely to be 
survivable than with 20,000.

	Unless you've been there and done that and definitely lost the 
job (or would have, if you had still been working there), you may not 
ever understand.

	You may not believe me, but I do hope that you never have to live 
through this kind of experience.

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(+++)

More information about the Mailman-Users mailing list