On 7 Jun 2016, at 16:04, Georg Lukas <[email protected]> wrote: > I think we can improve significantly for the XMPP IM use case (I have > not much expertise in other use cases), if we simplify the things that > can be simplified, make them consistent between clients, and give the > whole thing a catchy name like "Easy XMPP" or "Modern XMPP" or somesuch, > that will be applied to conforming implementations.
It’s worth noting that I think you mean the ‘Public/unaffiliated Internet XMPP IM use case”. Lots of XMPP use is either pre-provisioned or off-Internet, or both. Doesn’t detract from the need to do it, mind. > Many people come to XMPP because they want to chat to some existing XMPP > user. Therefore that user should have an easy mechanism to get their > friends onboarded, via some out-of-band mechanism (like a URL). That > onboarding should include client installation, account creation, adding > the first roster item, and optionally forwarding of "common buddies" > from the existing XMPP user: https://wiki.xmpp.org/web/Easy_Onboarding Apart from brief bits I disagree with (hiding password generation from the user), I think pretty much all of this can be achieved by a sufficiently motivated client already. > The above is a set of high-level ideas that apply to the whole process, > an in-depth look into a potential implementation (still far from a > proto-XEP) can be found in https://wiki.xmpp.org/web/Easy_Roster_Invitations > > The last one contains some open questions regarding the complexity of > the protocol and the possible entities where it should be implemented. > I'd like to have a discussion of these before I make my first attempt at > writing an XEP… I’ve only scanned it, but it doesn’t look immediately stupid on scanning, and I think writing up a XEP would be a good way to get further comment. /K _______________________________________________ Standards mailing list Info: https://mail.jabber.org/mailman/listinfo/standards Unsubscribe: [email protected] _______________________________________________
