[standards-jig] Packet Header JEP 0033

[Nicholas Perez]

I like the idea, but maybe we need to just define an attribute inside 
the x:header instead of people juggling three new namespaces.

just my US$0.02

Nick

Joe Hildebrand wrote:
> I've gone back and forth on the delivered='yes' attribute a couple of times.
> I tried capturing the delivery information in the trace elements, but that
> didn't work well.  After talking with Craig, the approach in the JEP was the
> one we decided would be the most implementable.  I guess I could imagine a
> couple of other ways of doing this, too.  Let me think about it some.
> 
> Another thing that I don't like is that the rules for delivery and
> re-delivery are a little complex, particularly with respect to BCC's.  It
> seems like there should be something simpler.  Maybe this is all part of the
> same problem...
> 
> Also, what do people think about splitting the three sections of the header
> into three different namepsaces, rather than putting them all into the
> header?
> 
> So, something like:
> jabber:x:multicast
> jabber:x:metadata
> jabber:x:trace
>