[Standards] Could anyone using XMPP-0277 or PEP help us test our Onesocialweb implementation ?
b+jabber at bruce-2010.zerlargal.org
Sun Apr 18 05:45:59 UTC 2010
On Sat, 17 Apr 2010, Stephen Pendleton wrote:
> My feeling is that XEP-0277 doesn't scale well, especially with PEP and
> frequently reconnecting mobile clients in the mix. In particular the
> rebroadcasting of messages (for example the replies to A that get
> broadcasted to everyone subscribed to A no matter if they are interested in
> that thread) creates a lot of traffic (and noise) on larger implementations.
> You can see what would happen with a large implementation.
As far as traffic goes, this isn't much diffferent from a busy MUC chat
room; the one-to-one (sending a post or reply) and one-to-many
(broadcasting the post/reply to subscribers) relationships are effectively
static compared to the number of messages/posts sent through. Presumably
large implementations would use the same solutions.
Regarding excessive noise caused by a deluge of replies on a given
item/thread, looking at XEP-0060, 16.6 could be used to unsubscribe from a
particular item in a given microblog, effectively muting that thread.
The particular wording reads:
| The "pubsub" querytype is defined herein for interaction with pubsub
| services, with three keys: (1) "action" (whose defined values are
| "subscribe", "unsubscribe", and "retrieve"), (2) "node" (to specify a
| pubsub node), and optionally "item" (to specify a particular item at a
from='juliet at capulet.lit'
to='paris at escalus.lit'
<item id='marry me juliet'/>
Using this to ignore a particular item may run into interoperability
issues with existing implementations however. I suspect that a number of
clients will effectively do this on the client side by simply not showing
any further replies to the user.
Huh, Count Paris doesn't seem to have a family affliation apart from
kinship with the Prince.
More information about the Standards