On Wed, Sep 1, 2010 at 8:33 PM, Justin Karneges
<[email protected]> wrote:
> On Wednesday 01 September 2010 12:18:22 you wrote:
>> On Wed, Sep 1, 2010 at 8:07 PM, Dave Cridland <[email protected]> wrote:
>> > On Wed Sep  1 19:39:58 2010, Kevin Smith wrote:
>> >> On Wed, Sep 1, 2010 at 7:38 PM, Justin Karneges
>> >>
>> >> <[email protected]> wrote:
>> >> > On Wednesday 01 September 2010 11:29:54 Kevin Smith wrote:
>> >> >> On Wed, Sep 1, 2010 at 7:29 PM, Dave Cridland <[email protected]>
>> >> >>
>> >> >> wrote:
>> >> >> > On Wed Sep  1 19:26:44 2010, Kevin Smith wrote:
>> >> >> >> On Wed, Sep 1, 2010 at 7:19 PM, Iñaki Baz Castillo <[email protected]>
>> >> >> >>
>> >> >> >> wrote:
>> >> >> >> > 2010/9/1 Dave Cridland <[email protected]>:
>> >> >> >> >> XEP-0145 provides a very simplistic capability for storing
>> >> >> >> >> notes about contacts on the server.
>> >> >> >> >>
>> >> >> >> >> Very rough, though - I'd be open to something better that
>> >> >> >> >> actually
>> >> >> >> >> tied into
>> >> >> >> >> the roster, personally.
>> >> >> >> >
>> >> >> >> > Yes, it seems very limited (just some text note). I would expect
>> >> >> >> > something as a vcard for each contact stored within the roster.
>> >> >> >>
>> >> >> >> In theory the server should be storing whatever child elements you
>> >> >> >> include in a roster set as well, although I don't know how many,
>> >> >> >> if any, currently do this.
>> >> >> >
>> >> >> > Gosh. Not us, but it does sound like the right thing to be doing.
>> >> >> >
>> >> >> > How does a client know this is possible, rather than using '49 or
>> >> >> > '223?
>> >> >>
>> >> >> Assume it is, because it's part of the RFC (or bis, anyway)? :)
>> >> >
>> >> > Erm, isn't there a risk that existing clients will destroy the
>> >> > metadata if the
>> >> > roster item is ever modified?
>> >>
>> >> There's a risk, but clients should be persisting content they don't
>> >> understand, rather than binning it.
>> >
>> > It does all seem to rely on luck and a following wind, doesn't it?
>> >
>> > Could we not make this a little more reliable, like roster versioning?
>>
>> We could - although it's not clear to me how much of a problem this is
>> - other storage specs suffer from it as well - Bookmarks for example
>> (either over private or P(E|O|I)P).
>
> I have a strong feeling no client stores arbitrary XML of roster items to keep
> around when (re)submitting changes.  At least not any designed before this
> became a topic.

Right, clients will need to be updated in assorted ways to match the
current bis drafts, I guess.

/K

Reply via email to