Developers of Sympa

[Jeff Abbott <address@concealed>]

On Oct 26, 2005, at 4:34 PM, address@concealed wrote:

> Jeff Abbott <address@concealed> recently wrote (in part):
>
> > Sympa, according to my understanding, uses the
> > $LISTNAME-request address to get in touch with the actual
> > person or persons running the list, the list owners.
>
> Just as a point of reference, this is also the behavior of
> ListProc, and I'm reasonably sure LISTSERV as well.  I'm not
> saying this is right and proper, but I will note that both
> these MLMs predate the RFC by quite a while.

That's good to know -- I've never used either so I'm not familiar  
with their details.

> Also I don't read the RFC to mean that the '-request' address
> is necessarily reaching a robot as verses a human.  In
> practice we've found that the '-request' is indeed used
> almost exclusively by subscribers sending free formed
> messages pleading to have their subscription removed.  :-)

Also valuable information.  I don't own any of our Majordomo lists  
that are active enough for me to really have to deal with user  
requests for a given list, and I do agree with you on the RFC being  
(intentionally?) vague on where mail sent to -request should end up.   
I suppose we'd have to decide amongst ourselves whether or not we  
thought the best practice was human or if having Sympa take -request  
mail would be better.

> > Logically, this is what I would expect $LISTNAME-owner to do,
> > instead of $LISTNAME-owner being used for bounce management.
>
> From the LISTSERV manual:
>
>    "If you find that you have users trying to contact you (as
>     list owner) at the owner-listname address, you should
>     tell them that the correct generic address for contacting
>     the list owner(s) is listname-request, not
>     owner-listname."
>
> Again, this is just historical precedent, and is not
> necessarily a good argument for what *should be*.

A big reason behind why I'm thinking about this is exactly because of  
historical precedent -- our historical precedent, thanks to  
Majordomo, is that users contact list owners via $LISTNAME-owner,  
though we do also have additional aliases for owner-$LISTNAME that  
point to the former for users who are accustomed to that format.  Not  
only would there be some retraining involved for an unknown number of  
users, but mail sent to $LISTNAME-owner seemingly disappears into the  
void if bouncequeue doesn't know how to handle it, so people's mail  
would be vanishing and neither the list owner nor the message's  
sender would be aware of it.  But the way things have been here might  
not have any bearing with the way things have been at other  
institutions, hence the robot-level configuration settings that would  
allow larger sites to more easily integrate Sympa into existing  
address standards.

> (BTW, I didn't see any recommendation in RFC 2142 for 'owner'
> but I might have missed it.)

There was no recommendation for -owner, so you didn't miss anything.  :)

> As for allowing the meanings of the addresses to be site
> specific, the only problem I can think of is having a Sympa
> installation out there that doesn't work quite like all the
> others.  But if the problems that causes for that particular
> site's users is far outweighed by the benefits seen, then I
> guess that should be that site's choice.

This certainly would be a potential source of confusion and trouble,  
but I'm of the opinion that it would be a potential problem for the  
listmaster, rather than confusing addresses that eat mail, which  
would be problematic for users.  Unfortunately there are still /way/  
more of them then there are of us, and they tend to whine a lot  
louder.  :P

> I can't speak to, however, to the additional work and
> complications it will cause in the overall development and
> maintenance of Sympa.

Hopefully that's where one of the core developers could chime in...

> Hope nobody minds my throwing in my 2 cents on this.

Absolutely not.  Knowledgeable input is always of value to me.

I know that I'm not the only one who's considered this since I've  
seen at least one reply sympathizing with my initial post, but is  
this solving the problem in the wrong way?  Should we at Duke just  
accept that users are going to be confused and annoyed for a while  
when their messages to $LISTNAME-owner don't get them in touch with  
the list owner (or, indeed, anyone or anything) due to the address  
changes?

Thanks,
Jeff

Attachment: smime.p7s
Description: S/MIME cryptographic signature