[Standards-JIG] Re: [jdev] XMPP Ping method?
kevin at kismith.co.uk
Thu Nov 2 11:52:00 UTC 2006
On 1 Nov 2006, at 19:31, Peter Saint-Andre wrote:
> Chris Mullins wrote:
>> That's quite a bit more complicated that I envisioned.
>> I was just thinking "IQ:Ping", and leave it there. For all of the use
>> cases I've come across, this would be sufficient.
(Responding to Chris here) This is sufficient for many things, but
the thing it doesn't do is the thing I'm most keen on achieving:
reliable delivery. I want to send a message and know that it'll
either be delivered, or I'll get an error.
> Probably it would be good to break the ping stuff out into a separate
> namespace and/or spec.
Is it worth a whole spec just for ping? As for the TLE vs namespace
stuff, I don't care too much, iq would seem to do fine to me, but a
ping being toplevel also has elegance.
> Do we need this to check a single stream only, or also to check the
> end-to-end connection?
If every stream is reliable, end-to-end reliability is also
guaranteed. So for acks, per-hop is sufficient. For pings, it's
useful to be able to do it end-to-end.
> IMHO we don't need stream feature negotiation to do pings, but we do
> need that for per-stanza acks (which are heavier).
I think this is correct.
Psi XMPP Client Project Leader (http://psi-im.org)
More information about the Standards