[petsc-dev] On "Reply-To" munging

[Satish Balay]

Sure 'reply-to:list' is not perfect.

But I'd rather keep this than deal with folks doing 'reply' - and
sending e-mails to individuals instead of doing 'reply-all - to list'.

Sure the 'reply-to-harmful' folks say: the user has the choice between
'reply' and 'reply-all' and 'reply-to:list' removes this choice [hence
harmful].

But I don't buy this argument. I think the default choice should be
the most-commmon used item. And for lists 'reply-all or reply-to:list'
should be this most-commmon use choice. But currently - this deault is
in the user's mail-clinet setting. Even if a very few users default to
'reply' instead of reply-all' there will be enough indvidual messages
to annoy us.

In my alternative scheme - 'reply-to:list' is the default [so it takes
care of the major usage]. In the minor usage where folks need to
*explicitly* replies to individuals - instead of the list - then folks
should do *extra work* and use 'forward' [instead of reply/reply-all]

In the case of 'David Sheehan' - he went back to replying to an old
petsc-dev email even though there were 30 new petsc-maint e-mails on
the thread. And based on the latest e-mail - he is not aware that he
is resending messages to petsc-dev and not to petsc-maint. In such
cases when the user doesn't know where he is sending e-mails, not
having 'reply-to:list' doesn't help anyway.

Satish

On Wed, 24 Mar 2010, Jed Brown wrote:

> This of course is the reason that David Sheehan keeps replying to
> petsc-dev, despite numerous requests to move it to petsc-maint.
> 
> The literature on why this is bad is plentiful.
> 
>   http://www.unicom.com/pw/reply-to-harmful.html
> 
>   http://woozle.org/~neale/papers/reply-to-still-harmful.html
> 
> I know many mailers are broken and don't have proper "reply-to-list"
> functionality.  I know this isn't the first time this has come up, and I
> understand that turning off the munging might cause more problems than
> it fixes, but I'd just like to point out the situation.
> 
> Jed
>