Re: [pgsql-www] Forums at postgresql.com.au

On Sat, Nov 20, 2010 at 05:46, Elliot Chance <elliotchance(at)gmail(dot)com> wrote:

> On 21/11/2010, at 12:32 AM, Magnus Hagander wrote:

>> On Sat, Nov 20, 2010 at 14:22, Elliot Chance <elliotchance(at)gmail(dot)com> wrote:

>> Use "Reply To All" when you want to send to the list. It's what
>> everybody else has been doing for ages :-) If you want to read up on
>> the bike-shedding that goes behind that preference, it is something
>> that comes up regularly - just search the archives.

>>> forums(at)postgresql(dot)com(dot)au is pointed to a black hole so that email disappears but the mailing list gets another copy. When the mailing list gets its copy it sends a copy to the forum (because the forum is just like a subscribed user), the parser then dissects the headers to find out where the post belongs. We already know this part works.

>> So how does one respond to the user?

> I can't explain it any clearer, your email response goes to the mailing list and that mailing list sends a copy to the original person thats how a mailing list works. It also sends a copy to the forum which is parses you and that person and anyone else can see the reply on the forum.

Elliot, Magnus wants forum->list email to come from a per-user address
so that when he replies directly to that address (without sending it
to the list), the response is mapped to a PM.

Magnus, I see a couple issues with that:
1 - Conventionally, private messages are not used quite that often in
forums. They are truly separate from public discussion, not a natural
part of it by virtue of being the same ultimate message destination.
2 - Since Reply to All is a convention on this list, the forum needs
to deal with that somehow. Both a private and public response
containing the same content is unacceptable for forums. The public
response is always preferred and can't be integrated later, such as in
the common "solution" of having the list software not send a copy when
it detects a person's address already in the address list.