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

Re: Mailing list headers [Was Re: ssh vs telnet - which is faster?]



On Fri, 10 Dec 1999, Steve Lamb wrote:

> Friday, December 10, 1999, 2:53:54 AM, Kenneth wrote:
> > I think that this is a bad suggestion since there are some people (such as
> > me) who use the Reply-To headers to ensure that the recipients have a valid
> > reply address to reply to us. If the mailing list rewrites the Reply-To by
> > putting itself in that header, then the valid reply address is lost.
> 
>     What prevents you from putting the valid address in the From: field
> instead?

	It is a perfectly valid address as per the mail server running on
my unconnected domain. RFC822 states that it should be authenticated, but
does not state who by. Read the section dealing with the Sender header. It
clearly allows for the case where the person that actually sends the
message is different from the author of the message. This is what is
important and ISP's authenticate on this header and not on the From header
(or if they are, it is unwarranted). The RFC indicates that the Sender
field was also intended to be used in the case of mailing lists which
redistributes email.

> 
>     Furthermore, it is an intented use of Reply-To as defined by RFC-822.
	
	Correct. But reread RFC822. This is only one of the intended uses
of the Reply-To header. The use of the Reply-To header to indicate that a
reply should goto another address because the From header address cannot
receive a reply is another one. In fact, given such an option, setting up
a mailing list to rewrite the Reply-To header is broken behaviour - unless
the intent of the list admin is to put such subscribers at a disadvantage.

	You may personally not like such use of headers. But if you want
to get the world to follow in your footsteps, write up a new RFC and get
it approved as a standard.

> 
> >         Furthermore, I think that the original problem of getting two
> > copies of the same mail can be solved easily by using procmail. 
> 
>     Assumption is that one is using or wants to use procmail.  The mail,
> however, is still sent.
> 
 
	That is true. However, since the RFC allows the behaviour I have
described above, you will have people following it. 'procmail' is a way of
handling the situation. If you dont want 'procmail', find another solution
or grin and bear it.

Regards,
Kenneth


Reply to: