Re: [Simple] SIMPLE and OMA and 3Gpp and RCS and… (new subject)

[Saúl Ibarra Corretgé <[e-mail]>]

Hi Ben,

On Nov 2, 2012, at 3:20 PM, Ben Campbell wrote:

> (as individual)
> 
> On Nov 2, 2012, at 7:27 AM, Iñaki Baz Castillo <[e-mail]> wrote:
> 
>>> I'd love to see a better model endorsed by SIMPLE and then happily implement it. I think we do share this goal :-)
>> 
>> Sure we agree here. The key is: reusing SIMPLE/XCAP specs? or making
>> something better from scratch (after learning from current SIMPLE/XMPP
>> specs)?
> 
> Do you really believe nothing is salvageable? For example, would you jettison all of the below?
> 
> -- Using SIP in the first place
> -- SIP Events
> -- RLS
> -- SIP Message
> -- MSRP
> -- XCAP
> -- etc
> 
> From the discussions so far, it sounds like most of the complaints are in the area of reusing user lists between clients (e.g. contact lists, presence rules, etc). Do people really hate the entire SIMPLE suite, or is it just an XCAP issue.

IM and SIP events are nice, the problem I've experienced with SIMPLE was letter P. The fact that a complex protocol (XCAP) was used along with multiple documents, and stretching resource-lists to contain a buddylist is what creates trouble. Actually XCAP is fine for storage, if we remove that default namespace problem and find a way to have something like xcap-diff which contains full contacts every time.

> 
> If the scope of the work is "start over from scratch", then it seems like we would be better served starting with XMPP. I'd be very surprised if we could get support for creating "yet another presence and IM protocol".
> 
> We've been there before and it wasn't pretty.
> 

Not sure if anybody has had problems with the IM part, but I'm pretty sure Iñaki is talking just a bout the presence part, just as I am.

Now, assuming we are just considering presence here, we could go either way: create a completely new spec which reuses nothing from the current (or maybe SIP events) or try to reuse what would "make sense" without compromising the whole design. The problem is that if we go the reusing route, is it acceptable that all specified documents are discarded,a  new document type is created and XCAP default namespace handling is fixed? I would not be backward compatible, of course.

IMHO this is the key point before starting to design any solution to the problem, because if it won't be accepted / adopted no matter how clever it is, we better do something else.


My 2 cents. Regards,

--
Saúl Ibarra Corretgé
AG Projects