> Please reread the numerous debates held on this topic in the past. I will
> not comment on anything that has already been said.
I don't need to reread them, they were largely between you and I. :-)
> >ALSO, "Reply-To:" would not be necessary, because a simple REPLY command
> >would then reply to the sender,
>
> Which, in your example, points to an error-delivery mailbox.
No. Poor wording on my part. By sender, I meant the person, whose
address appears in the "From:" line. Doesn't RFC822 say not to have
the REPLY command ever automatically use the address in the "Sender:"
line?
> >and REPLY ALL would reply to the list.
>
> and to the error-delivery mailbox, or, assuming there was no "Sender:",
No. Explained above.
> to the list and the individual who originated the message, who would then
> get 2 copies.
I consider that an extremely small price to pay for the other
advantages which would result.
> >This would solve the annoying problem of postings stating "reply to me"
> >where the sender doesn't insert a "Reply-To:",
>
> As you said yourself, this annoying problem can be solved by the person
> typing "Reply-To: <his userid>" rather than "please reply to me as I am
> not subscribed to any of the lists I am sending this message to", which
> is much shorter and also more efficient.
True, except for the fact that the vast majority of people on LISTSERV
lists these days are not sophisticated network users who know about
these things. And even some sophisticated people forget now and then.
> >And since everyone would get an identical copy of each posting, the
> >DISTx code might be much simpler, and it might also be much easier to
> >save more network bandwidth.
>
> Please elaborate on this. I'd be interested to know how this would reduce
> the load, given that the mail headers would then reach an average of 30
> lines rather than the usual Date/From/Sender/To/Subject that LISTSERV
> presently generates.
Because rather than one copy for each 5 or fewer recipients per node,
one copy period would be sent to each node. I recognize that this is
not normally going to be significant, but it is not a detriment either.
And why would there be any more lines in the header? The way I count
it, if you remove "Reply-To:", that's one line less, not n lines more.
> And oh, by the way, you don't need LISTSERV to
> implement a list like that. The Crosswell mailer will do it just fine,
> and more efficiently.
>
> Eric
Yep. Most mailers can implement lists. So, why do people use
LISTSERV lists? User-initiated subscription/unsubscription, so that
it doesn't occupy the listowner's time. The other consideration,
which fewer people are concerned about, even though it's more
significant (on Bitnet) is the savings in bandwidth from DISTx. Every
mail expert that I've discussed this with says that LISTSERV's use of
"Sender:" is in violation of RFC822. I like LISTSERV. I do however
think that the "To:" and "Sender:" field usage is not good. When I
receive a posting from a list, that posting is not "To" me, it's "To"
the list. I find "To: ldw" distracting.
What about the bandwidth that's wasted when some new and slightly
different rejection notice slips through LISTSERV's filters and gets
distributed to the list?
Some day maybe I'll give up on this argument, since everyone else
(including those who agree with me) has advised me that your mind is
made up, and that you will never change this.
/Leonard
|