Yes, PubSub is definitely full of win. Interestingly, the XMPP/AtomPub
bridge is built on PubSub, so all the pieces are *almost* there:

> http://www.cestari.info/2008/6/19/atom-pubsub-module-for-ejabberd

Re-iterating my previous statement, perhaps we should approach the
author about filling out the functionality some and supporting it in
some way?

--Steve

On Thu, Jun 19, 2008 at 2:27 PM, Jean-Marc Liotier <[EMAIL PROTECTED]> wrote:
> Also sprach Steve Ivy [Thu, Jun 19, 2008 at 12:20:23PM -0700] :
>>
>> At which JID should I receive these notifications? I don't really want
>> "so and so would like to friend you" and "your friend is now a zombie
>> and ate your brain" messages coming to my iChat, adium, or gaim
>> window. And yet, I don't really want to maintain any more JIDs than I
>> need to.
>
> That is what publish-subscribe is about : you only susbscribe to the
> streams that you want to receive notification from.
>
> And then not all messages need to be rendered as chat on the client
> side. For now most XMPP clients are in fact chat clients - and the
> current user experience of notifications through XMPP is therefore the
> reception of a message from a "contact". But those notifications could
> also be rendered as a news stream the way we render Atom or RSS, or it
> could be a ticker tape, or a tray notification popup or whatever you
> want to develop your client as...
>
>> In addition, there's the issue of message capture by different
>> clients. If I *am* logged in via a chat client to the same JID that my
>> DiSo-enabled site is using, will I lose messages if the real-time
>> client is available to receive it but my "social inbox" client has not
>> connected recently?
>
> When server side history becomes widespread (probably through
> implementations of XEP-0136) this problem will cease to exist just as
> all my mail clients see the same messages on my IMAP server.
>
>



-- 
Steve Ivy
http://redmonk.net // http://diso-project.org
This email is: [ ] bloggable [x] ask first [ ] private

Reply via email to