[Standards] Move Carbons to Last Call ("Proposed")

Georg Lukas georg at op-co.de
Mon Aug 17 17:37:37 UTC 2015


* Kevin Smith <kevin.smith at isode.com> [2015-08-17 15:47]:
> After discussion in the XSF MUC, I’ve pushed another change, so
> probably best to track via the
> https://github.com/Kev/xeps/commits/carbons branch.

I'll see your patch and raise you $200.

https://github.com/ge0rg/xeps/tree/carbons

After some more discussion in xsf@ we figured out that at least two
server implementations (prosody, ejabberd) haven't implemented section 6
of XEP-0280 anyway, and were using identical behavior for full-JID and
bare-JID messages. Furthermore, section 6 is completely redundant, as
the same behavior can be achieved under existing RFC 6121 §8.5.2.1.1
rules.

OTOH, message forking (as opposed to carbon-copying) introduces several
"problems":

 * a carbons-enabled client with a negative priority can receive chat
   messages under the section 6 rules.
 * a carbons-enabled client can not determine if it received a bare-JID
   message due to regular message routing or due to Carbons (this is
   useful for determining if a notification should be silent or loud,
   though a "smart" client can determine that by watching presence of
   the user's other resources).
 * it makes the XEP more complicated for no benefit.

Therefore, I have gone a step further than Kev and removed section 6
(and rephrased section 7 accordingly). As this change does not break
anything, I'd like to have it added to XEP-0280. However, it is based on
Kev's patches, so please discuss it now, and I'll open a PR as soon as
Kev's changes are (hopefully) accepted.


Georg

P.S: This was already discussed a year ago, the thread can be found here:
http://mail.jabber.org/pipermail/standards/2014-April/thread.html#28807
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 811 bytes
Desc: Digital signature
URL: <http://mail.jabber.org/pipermail/standards/attachments/20150817/faa8ee92/attachment.sig>


More information about the Standards mailing list