Oh, my list is missing the carbons of course. 2015-04-18 10:58 GMT+02:00 Stefan Strigler <stefan.strig...@gmail.com>:
> Sounds like a really nice hack. A recombination of presence, disco and MAM > to gain a totally different user experience. > > +1 for the idea :) > > Not sure where to put this though. How about > > XEP-1337 Hacks > > :D > > 2015-04-18 5:24 GMT+02:00 Kurt Zeilenga <kurt.zeile...@isode.com>: > >> >> >> > On Apr 17, 2015, at 7:57 PM, Peter Saint-Andre - &yet <pe...@andyet.net> >> wrote: >> > >> > The Message Archive Management spec (XEP-0313) seems to assume that a >> message archive will live on the server where a user has registered an >> account. This raises privacy and security concerns, especially if the >> messages are not encrypted: as a user I might not want all that message >> history on the server in case it gets hacked, and as a server admin I might >> not want the liability of holding all those messages, either. (In fact, as >> someone who runs a very large public IM service, I can assure you that I do >> not want to have all those messages entrusted to me!) >> > >> > Ideally, to me, my message archive would be stored on a trusted device >> that is under my control (say, a limited-access storage medium that I keep >> in my house). This device could authenticate to my account and advertise >> its existence to my other resources. Using Carbons (XEP-0280) it could >> obtain copies of all the messages I send and receive. When one of my >> messaging devices wants to retrieve message history, it would do so by >> querying this trusted storage device, not the server (which only handles >> messages for purposes of realtime delivery). >> > >> > I would really like to see the wording in XEP-0313 adjusted to take >> this scenario into account. I am happy to propose text. >> >> I think MAM should be mostly accessing server maintained archives. If >> the archives are maintained by some other entity, such as a client under >> the control of a user, some other extension is needed to address the >> particulars of this scenario. For instance, discovery (the advertisement >> you noted above) would be completely different. I rather not attempt to >> detail this scenario in XEP 313. I don’t see any particular need to change >> XEP 313 text to enable a client to offer MAM services. I think that’s >> already allowed. For instance, Section 7 says “If a server or other entity >> hosts archives and supports MAM queriers…”. >> >> — Kurt >> >> > >> > Peter >> > >> > -- >> > Peter Saint-Andre >> > https://andyet.com/ >> >> >