Collecting requirements for a KDE-wide instant messaging solution (was: Re: radical proposal: move IRC to Rocket.Chat)

[Elvis Angelaccio]

On mercoledì 9 agosto 2017 00:19:32 CEST, Thomas Pfeiffer wrote:
> Hi everyone,
> now that hopefully most of the emotional arguments in fiery support of one 
> protocol or another have been exchanged, I'd suggest we move 
> things towards a 
> practical approach and ask ourselves:
>
> What are the requirements that KDE has for an instant messaging 
> / chat system 
> for it to be viable as our main channel for real-time communication for the 
> foreseeable future?
>
> Here is what I could come up with, feel free to add new requirements or 
> challenge the ones I'm listing.
>
> Must-have:
>
> - FOSS clients or at least API available for desktop as well as mobile
> These clients must 
>  - have a UI that someone who is < 20 years old and cares about 
> the looks of a 
> UI would use (or if those don't exist, we need to have people 
> willing and able 
> to write them before switching) 
>  - run smoothly on computers that can run most other KDE software, without 
> eating all of their memory
>
> - FOSS server implementation
> (this might look like a nice-to-have for some, but if we'd 
> require everyone in 
> KDE to use it, it's not optional)
>
> - Ability to use without having to create a new account just for that.
> We could force contributors to sign up for something, but we'd increase the 
> barrier of entry if we'd make it mandatory for everyone who's just curious 
> about what's happening in KDE.
> Identity would suffice, as everyone who does anything with KDE 
> has an Identity 
> account anyway.
>
> - Permanent logs across mobile and desktop clients without the 
> need for users 
> to set up anything.
> That means ZNC does not count unless we implement it in a 
> desktop as well as 
> mobile client in a way that is completely friction-free for users
>
> - Easy way to share files
> A solution that puts files automatically on share.kde.org and 
> embeds them from 
> there works only if we have people willing and able to 
> implement that feature 
> into a desktop- as well as mobile client
>
> - Support for a decent set of Emoji (not just the ones you can create using 
> ASCII chars).
> Using Unicode to display them is probably okay, as long as users can choose 
> them from a menu in the client instead of having to paste them from 
> KCharSelect.
> This, too, might sound like nice-to-have for many, but not 
> having them would 
> cut us off from the younger generation. Yes, they use them even in a 
> "professional context". Believe me, I'm seeing it in action 
> every day at work.
>
> - User avatars
> Again, must-have if we want to reach the younger generation
>
> - Uses a port that is open even on educational networks
>
> - Channel listing
> So that every public channel can be easily found
>
>
> Nice-to-haves:
>
> - Bridge to IRC
> For the transitional period or for people who just refuse to change their 
> habits
>
> - Full name display
> Makes things feel more trustworthy
>
> - Integration with our development tools such as Phabricator
>
> - Web client
> Very handy if you are at a device which isn't yours and quickly 
> want to check 
> up on things
>
> - Stickers
> People love them when they have them, but they survive without them.
>
> ---
> I'm sure I've forgot many things, but this (already quite long) list should 
> give us a good start.

We should probably also ask the sysadmin team whether they would be willing 
to maintain our own chat server. Currently freenode is a third-party 
service, while switching to a self-hosted solution would probably increase 
the burden on the sysadmin people and also on the server infrastructure.

>
> Looking forward to a productive discussion,
> Thomas
>
>