On Sun, 21 Apr 2019 12:33:21 +0200 Christian Boltz <opensuse@cboltz.de> wrote:
[A lot of administrative questions, if openSUSE moves to mailbox.org]
Answering all these questions might be a lot of work - and you might end up with not implementing it at all if the majority of your users is either against it or abstain from it.
Thanks for playing [1] devil's advocate ;-) and all these questions.
I'll give you a collective answer:
Now we wanted to tell our members about it before we spend (or possibly waste, if nobody wants it) time on it.
Very good idea! Full support for this openess! Truely!
This also means that there are still some open questions we'll have to check and answer.
Absolutely understandable - and I just wanted to add the basic questions above for some "impressions" about what needs to be clarified. Especially with GDPR and all latest changes, I salute you to your courage to open this can of worms! :-)
To name a few things (note that this reflects my personal opinion because we didn't discuss these details in the board or heroes yet) - aliases will stay the default, the mailbox should be opt-in (everything else would lead to "dead mailboxes" with unread-forever mails) - Heinlein / mailbox.org should become the primary mail server for opensuse.org (= DNS MX entry) because that would give us better spam filtering/blocking (sorry to say that, but the spam filter on mx*.suse.de is far from perfect)
Do you have some comparable numbers? How much spam is detected in a better way by Heinlein than by mx.suse.de? Could it be that mx.suse.de is only tagging spam and not deleting it because of legal rules? ...could it be that admin@o.o gets some special handling because of some historic settings?
- obviously, this also means Heinlein will need/get a list of all @opensuse.org mail addresses and their alias target.
Why does this need to happen? If I remember correctly, Heroes complaint about the complexity of the openSUSE mail system already (because it is connected to the SUSE development/testing mailsystem) - now you want to move it to another company. How does this make things easier? I would in turn recommend to start with the final separation of the openSUSE infrastructure from SUSE: 1) setup/register your own offical DNS servers (as you have the internal ones already, this should not be that complicated) 2) setup own MX server (you can clone the mx.suse.de ones, if you like) 3) speak with MF-IT about the forums, blogs and authentication stuff => each of these steps is independent. But with 1 & 2 the openSUSE heroes would have the full flexibility to look at the Mailsystem or any other new service in their timeframe and with their power.
(In theory we could avoid giving them the alias targets, but that would mean to a) not have their spam filter or b) to continue to have the aliases on mx*.suse.de and relaying the mails over it or c) setting up a mailserver in the openSUSE infrastructure that does the alias forwarding. I have to admit that I dislike all of these options because they add an additional and IMHO superfluous step.)
Well, with aliases as default, mailing lists and administrative accounts (like {post,web,mail}master@o.o, admin-auto@o.o, ...) and all the different (sub-)domains (like opensuse.de for example), I'm really not sure if your solution is really the more easy one to implement. I instead would vote for c, as my former comment already implied. And I guess I'm not alone with these "independence" ambitions, if I read the other mails in this thread. In turn, I'm wondering how this matches: while some board members want more independence from SUSE, you try to push an important communication channel under a new umbrella - to a company that has not much to do with openSUSE at all.
(Do we get more members? Does Heinlein get more users/user data for marketing?)
I know Per Heinlein personally, and I'm quite sure he or his company does not look for new "marketing victims" ;-)
I do the same. But what happens if Heinlein get's aquired by another company or Peer (yes: two 'e') steps back from his position?
And no, I don't really expect that we get more members just because we start to offer mailboxes.
So this is really just something that someone - who? - wants to offer the existing members.
2) How many members currently complain about "missing mailboxes" at all? (all that paperwork for just one member?)
I remember several people asked for real mailboxes and/or being able to send mails using the o.o address and/or reported problems with the alias like SPF fun. (Sorry, no exact numbers. There's clearly more than one, but I'm too lazy to go through the ticket archive to count them ;-)
So let's wait for the amount of feedback to your request and wait with further discussions until we reached a deadline (that you should still define)?
4) Who will implement the needed scripts or do the manual work to manage the members mailboxes? (and does this lucky guy already know about this?)
Another "detail" we didn't discuss yet - but if needed, I'm willing to help with it. Having better spam blocking (in my case I'm especially interested in admin@o.o) is more than worth it, and more fun than deleting ticket spam all the time ;-)
What I take from this: From your point of view, the main reason to move all member Email accounts (as this is what you told earlier with "Heinlein will manage the aliases) to Heinlein is a better spam filtering for one Email address named admin@o.o. Right? Trying to summarize Benefits when moving from SUSE to Heinlein: * better spam tagging, done by Heinlein who get all @o.o Emails * possibility to sent with an <alias>@o.o Email address * openSUSE could blame Heinlein if something is broken * Heinlein could blame the Heroes if something is broken (heya: ping pong) Benefits when providing an own MX: * spam tagging can be done with openSUSE packages on openSUSE machines * everything else can be done with openSUSE packages on openSUSE machines (someone might even create some docu around it?) * members could be able to use their account data to sent with <alias>@o.o via the new MX after authentication * openSUSE could only blame the overloaded Heroes (or packagers) if something is broken In any of the two options, the current setup (including connect.o.o) needs to be adjusted.
[1] I know you - and wonder if I should say "being devil's advocate" instead ;-)
I hope I did my job right with my answers above ;-)) With kind regards, Lars -- To unsubscribe, e-mail: opensuse-project+unsubscribe@opensuse.org To contact the owner, email: opensuse-project+owner@opensuse.org