On Mon, Oct 17, 2016 at 12:54 PM, Jonathan Riddell <j...@jriddell.org> wrote:
> <Sho_> * Services integration - I see more and more users who have > data locked up in clouds (GDrive, etc.) and rely on accessing it > <Sho_> Gnome has GDrive integration, we need to do more on getting > cloud filesystems into our software too, but we dropped the ball on > KAccounts etc. for now > As the former KAccounts maintainer, I'd like to share some insights on this. The tech is there. It's proven by real world usage in Ubuntu (incl. Phone), Meego before that (iirc) and to some extent also Gnome, which has its own fork of the same thing (I guess they just really don't like Qt anywhere). Also Elementary is using one of these, but I don't remember which. And if I'm not mistaken, also Jolla in their phones. So we have the tech, but what we're missing are two things, really. 1) Things actually using it. We've had KAccounts for a while but there was nothing taking advantage of it, except KDE Telepathy, which migrated fully to it. So it sort of became a synonym for KTp accounts config. On the other hand, there was nothing much else that would actually need it. There was an OwnCloud lib or Plasma thingy at some point where I pleaded to have the account setup done through (K)Accounts to boost the adoption, but it never happened. Related to that is... 2) Buy in from PIM. I mean, let's face it, the real accounts user in the system is PIM, with all the emails and calendars and contacts and what not. This obviously requires quite some large effort to actually migrate everything to it and sadly the PIM team is long time understaffed. Now with the new Sink thing, this could have been a chance, but, again, after asking couple times about the accounts plan, (K)Accounts was never in the picture. And that's a problem. And so now we have this system that nobody in/around KDE wants to use. My take on it is that people (devs) don't really understand how that system works and how to take advantage of it, which is partly my fault. It is fairly complex system on the inside with lots of points of possible breakages, but also allows for great flexibility all around. So, imho, in order to successfully follow literally every other desktop these days, it needs a proper full integration in all things using accounts. And it really shouldn't be just an afterthought. It needs to be put right at the core of those things and workflows should be redesigned to take that into account (heh). It needs to be a conscious community-wide effort, like Frameworks was. Otherwise KAccounts will stay exactly where it is now - a promising (egg-)shell that has no real use. Unfortunately if the past two years are any indication, especially with Sink, the road ahead is not very bright. I'm happy to try my best to answer any possible future questions about the system you might have. Friedrich Kossebau was also writing a tutorial about it at some point, but I'm not sure where it ended up (if at all; I also failed to follow up on that). Cheers -- Martin Klapetek
