Re: SIMPLE and OMA and 3Gpp and RCS and… (new subject)
Saúl Ibarra Corretgé <[email protected]> Fri, 2 Nov 2012 13:25:42 +0100
| Newsgroups | gmane.ietf.simple |
|---|---|
| Message-ID | <[email protected]> |
On Nov 2, 2012, at 1:07 PM, I=F1aki Baz Castillo wrote: > 2012/11/2 Sa=FAl Ibarra Corretg=E9 <[email protected]>: > = >>> buddylist: >>> -------------------- >>> sip:[email protected] { >>> presence-subscribe-to: true, >>> presence-allowed: true, >>> presence-blocked: false >>> } >>> = >>> sip:[email protected] { >>> presence-subscribed-to: true, >>> presence-allowed: true, >>> presence-blocked: true >>> } >>> -------------------- >>> = >>> And that's all. This is easy to render for the watcher and easy to >>> process for the server. Single "document" with buddies and their >>> attributes. >>> = >>> No need for "external references to other XCAP documents in any other >>> server in the world", no need for generating a coredump if the same >>> buddy is contained in "oma_my_buddies" list and "my_blocked_contacts" >>> list and "oma_PoC_contacts" list and "oma_featured_buddies" list. >>> = >> = >> This is what we did in our addressbook implementation, while maintaining= compatibility to the highest possible degree. > = > And what is the point in "mantaining compatibility" if no other SIMPLE > client will understand the advanced document format of your > resource-list document? (they will probably delete it and regenerate > it). > = > And, if we just need a single document (based on buddies with > attributes) why do we need so many scattered XML/XCAP documents in the > server? and why does the presence server need to deal with N XML > documents and external references between them? > = > Why do we need a "RLS document" for presence subscription if we can > set an attribute "presence-subscribe-to" for each buddy in our > buddylist so the server already knows which users to subscribe to? Why > do we need a pres-rules documents if authorization rules are given via > buddy attributes? why do we need 95% of SIMPLE/XCAP specs if a single > document is enough? > = > And why to mantain "backward compatibility" if there is NO one full > working SIMPLE client out of walled gardens (apart from Blink which, > finally, works well but just when using the server infrastructure > designed for it)? > = > = You misundertood. That how we did it *for now*. I never ever mentioned ther= e shouldn't be a better way nor that this is how I (personally) would like = SIMPLE presence to look like. But I can't just simply implement a solution = which only works with myself. Hence the backwards compatibility. I'd love to see a better model endorsed by SIMPLE and then happily implemen= t it. I think we do share this goal :-) -- Sa=FAl Ibarra Corretg=E9 AG Projects