Re: Annoying Reply-To - Mailing list pgsql-general

On Fri, 17 Oct 2008 08:56:34 -0400
Stephen Frost <sfrost@snowman.net> wrote:

> * Bill Moran (wmoran@collaborativefusion.com) wrote:
> > You can resent it or not, but this _is_ a personal thing.  It's
> > personal because you are the only one complaining about it.
> > Despite the large number of people on this list, I don't see
> > anyone jumping in to defend you.
>
> Ugh.  No one else is jumping in simply because we've already been
> through all of this and it hasn't and isn't going to change.  The
> PG lists are the odd ones out here, not the other way around, I
> assure you. One might compare it to our continued use of CVS.
> It's wrong and backwards and all that, but most of the PG
> community is used to it and changing is a pain.  shrug.

I'd say because postgresql list has been used to it by a longer time
than most of the new comers doing the other way around did. But it
seems that the new comers are the most vocal.
Maybe because what's in the email headers have been abstracted to
them long ago they never had the need to look what they are for and
use them properly getting all the functionality their way has and
some additional bits too.

It is surprising how many people think to have enough knowledge in
email distribution systems to discuss an RFC that has been
already rewritten 1 time.
It would be nice if all those people spent their time rewriting in a
coherent way that RFC so that Reply-To works as they think is best
for the overall Internet without breaking any already existent
functionality before challenging this list consolidated habits.

Settings of ml are generally a mirror of their community.

Decent email clients have a switch for reckless people. That's
freedom.
Mine is called: "Reply button invokes mailing list reply".
Decent lists generally have quite helpful headers for filtering and
choosing to reply to the right address.

If you're for freedom... then let the recipients choose. Not the
list. If people insist in badly configuring/choosing their email
clients how far are you willing to go to protect them against
themselves and imposing a toll on the rest of the others?

I think the overall amount of the time I spent choosing the right
button in my life is lower than the time a single person has spent
writing 1 post on this topic and much much much lower than the time
they will have to spend in excuses (if anything worse) the first
time they will send to the wrong address.
But still it is much more than the time the people are complaining
have spent reading RFC 2822 and considering its implications.

But maybe this will give everyone a chance to consider all the small
coherent technical details that good engineers placed deciding about
email headers and email clients and reconsider what RFC are there
for.
If headers are properly set the action taken once you press your
chosen button is unambiguous and you conserve as much information as
possible.

If you think the majority is right since most of the people that
arrived to the Internet late got used to mangled Reply-To I think
that mistakes are educating. But till people will ignore what's
available and why I bet they will just learn to wait 20 minutes
before sending any email after their first expensive error, rather
than considering other ways to operate.

BTW consider this even from a HCI point of view... you still need 2
functions: one to send to "list" one to "author".
Saying you just want one button since 99% of times you'll reply to
the list is making the wrong expensive choice even more probable.
Once you've "2 buttons" why should you mangle the headers and give
them meaning they don't have?
Because most of the people aren't able to properly chose and
configure their email client?
What's wrong between making the difference in the header, in the
client and in your mind between Reply-to and List-Post?
Or is it better that the client thinks that Reply-To is replying to
the Sender (that's not true), the mailing list thinks that the
Reply-To is the List-Post (that's not true) and you think that today
is a good day to play lottery (In Italy it may be true)?

If you're among the reckless people you could get used to invoke the
"send to list" button and let your client send it to the Reply-To in
case there is no List-Post... or whatever bad habit you enjoy more.
But I see no reason to harass an RFC.

What about non standard web sites that have to cope with not
standard browsers and then being forced to adapt browsers that were
already standard to cope with non standard web sites?
Does it sound familiar?

I beg everyone pardon, especially to Tom Lane whose replies always
shine here, but I couldn't resist to reply to people thinking I'm not
"sensible", I took it personally ;) Evidently I'm old enough to
know the existence of RFCs but not mature enough ;)


--
Ivan Sergio Borgonovo
http://www.webthatworks.it