[Standards] Regarding XEP-0166

Peter Saint-Andre stpeter at stpeter.im
Mon Jul 9 15:17:07 UTC 2012

On 7/9/12 9:03 AM, Todd Herman wrote:
> I am working on updating my C# XMPP library to support  XEP-0176.  Since
> I don’t know too much about the main subject I first read through the
> STUN and TURN specifications.  I am currently finishing up reading
> through the ICE specifications now and will be starting on XEP-0176 very
> soon.

Thanks for your interest. Are you talking only about XEP-0166 (the core
Jingle spec), XEP-0176, or also other specs in the Jingle suite?

> The main point of this message is that I will most likely have many
> questions related to the XEP in the very near future. 

I think it is best to ask such questions on the jingle at xmpp.org list:


> I also noticed
> that it is draft and wanted to know what I could do to perhaps push the
> extension to the next step and take it out of draft.

We're not always in a hurry about pushing specs from Draft to Final. The
best example is XEP-0045, which is widely implemented but has been Draft
since 2002. Another is XEP-0047 (In-Band Bytestreams), which we just
recently pushed to Final but which had been Draft since 2003.

So you can see that we like specifications to be stable and widely
implemented before we go from Draft to Final. That said, if a technology
meets those conditions then by all means let's consider it.

Is there a particular hurry in the case of XEP-0176?

I'll also note that the fact that you have many questions about XEP-0176
might indicate that it's not quite ready to go to Final. :)

> Also, I do have one specific related question.  Both the ICE
> specification and XEP-0176 mention that candidates are validated by
> sending STUN binding requests and responses.  The specifications
> indicate that these responses and requests are sent between the clients
> (assuming that is what the candidates refer to) rather than to a STUN
> server.  I wanted to confirm that this is accurate and that this means
> the clients, which are XMPP clients, would also need to be able to
> function as STUN servers in order to interpret and respond to the requests.

I'll forward that question to the jingle@ list for discussion there.


Peter Saint-Andre

More information about the Standards mailing list