On 2014-06-23 20:25, Ivan Vučica wrote:
> I've taken a few moments and looked more closely at XEP-0313. A few
> comments follow, primarily about section "5.1 Simple configuration". 
> 
> This section can be interpreted as if settings for all jids have to be
> retransmitted, even if only one change has been made. Doesn't that
> introduce a possible race condition in case two clients attempt to make
> an update at the same time? Doesn't that unnecessarily make the
> transmitted stanza (or a received confirmation of the new preferences)
> bloated?

This has been discussed at some point already.  IIRC, the consensus was
against doing anything about that.  You would need to get preferences
before changing them.  I don't think they would normally be changed at
the same time from multiple clients, as that should probably be a
user-initiated action.
> 
> Is there a reason for the syntax <always><jid>#textnode</jid></always>?
> Why not, for example, <always jid="#jidhere"/>? Can <always/> and
> <never/> repeat in the first place? If so, can there be more than one
> <jid/> in each instance of <always/> and <never/>?

You include one <jid/> element per jid under the <always> or <never>
element.

> 
> A server implementation can get around a lot of issues around
> configuration by exposing ad-hoc commands. But, that doesn't let a
> client design a nice UI. And, how does an archive configured using
> ad-hoc commands map to simple configuration?

A basic form might be something to standadise, probably with 'always'
and 'never' as jid-multi fields and a 'default' field with the options
'always', 'never' and 'roster'.
> 
> How does one read the simple preferences without modifying them? With
> <iq type="get"/>? If so, can that be explicitly stated in the text of
> the XEP?

<iq type="get"> should never modify state.


--
Kim "Zash" Alvefur

Attachment: signature.asc
Description: OpenPGP digital signature

Reply via email to