Here and also in the github issues, there are many request for switching / upgrading / adding / enhancing the web clients we currently support. Roundcube and owncloud that is for now. I think it would be much cleaner, if MIAB would drop support for web clients completely and rather focus on the mail/contacts/calendar server part. Moving to radicale and dropping owncloud, would be one huge step in that direction.
Of course, we could provide some guides on how to connect to MIAB using your own owncloud instance. Maybe there is even room for another project that provides easy installation for web clients.
The big benefit would be, that the development can focus on the important parts, like stability of the mail server, security and testing.
Well, there are so many good desktop and mobile clients available. I would guess, that these clients are the primary access point for many users already. And it seems many users actually using the provided clients, aren’t really happy about them either.
My point is, that making a clear separation between backend and frontend would be benefitial. For me, the back-end part is the core of MIAB. frontends are just add-ons, that could be installed on top of a MIAB server (or even better on a different one)
ok, it probably makes sense to have a simple web mail client available. However, I think MIAB should limit the UI options to be simple and easy to maintain. Things like owncloud or sogo or any other complex solution shouldn’t be part of MIAB. If someone wants something more advanced or fancy, there is always the option of installing it yourself.
I like this idea. But: Then we should really think about some plugin system for frontend things like owncloud / webmail etc. Then we would have the MIAB core that solely provides server components like mail, logging, ips, backups that itself manages plugins for the frontend components. This would allow users to adjust their instance based on their personal taste (which would in fact end the discussion about ui components immediatly). On the other hand, users (like me) are still able to have one server for ui and backend. Forcing users to rent a second server in order to host their webmail etc. is a very bad idea, at least from my point of view.
I also think it’s not that hard to implement such a plugin system. Plugins could be github repos that follow a certain structure. That means, they provide install / backup / migration-scripts at fixed places. Maybe they are also assigned a data location by MIAB that is then backed up by MIAB within the schedule. Of course we would have to think a bit about a good API for the plugins, but I’m certain it’s possible and will take away some workload from the core developers while enhancing the usability at the same time.
If a lonely user voice means anything - please keep a reasonably competent webmail frontend…
I actually use the MIAB Roundcube fontend everyday and find it competent enough. The killer feature for me is the ability to change the Reply-To address on the fly for each email. I have searched and not found an Android client which can do the same as easily. Also with Mozilla dropping their mail client - a good webmail frontend is becoming essential.
The plugin idea is really neat and I really think we should aim for something in that direction. The first step would be separating and untangling front and backend parts. And this step makes sense for me no matter what direction we go. The next step would be making the installation of all front-ends consistent. That could be something along the lines to what ma0ho describes: scripts at the same locations to install / backup / migrate. Once we are at that point, we should have a good idea what it takes to plug in front-ends and could allow adding different ones.
And yes, a simple UI should always be part of miab. But I think we need some strategy to answer the never-ending discussion about UI components.