Re: Does 6122bis need to update 5122?

On 8/8/14, 2:27 PM, Ben Campbell wrote:
> Did Peter's comments go to the list? I did not see the original.

I never replied in this thread. :-)

> One comment below:
>
>
> On Aug 1, 2014, at 11:55 AM, Joe Hildebrand (jhildebr)
> <[email protected]> wrote:
>
>> On 8/1/14, 10:49 AM, "Peter Saint-Andre" <[email protected]>
>> wrote:
>>
>>> In that case, we'd specify an XMPP URI (RFC 5122).
>>>
>>> Thus I might suggest:
>>>
>>> For an XMPP client [RFC6120] using STUN and TURN, the ORIGIN
>>> attribute is an XMPP URI [RFC5122] representing the domainpart of
>>> the client's Jabber ID (JID) [RFC6122]; for example, if the
>>> client's JID is "[email protected]/balcony" then the ORIGIN
>>> attribute would be "xmpp:im.example.com".
>>
>> This made me wonder if we need to either rev 5122 to point to
>> 6122bis, or just have 6122 update 5122 (which I think I'd prefer,
>> so we don't open up a can of IRI worms).

I see this chain:

5122 normatively references 3920
6120 obsoletes 3920
6120 normatively references 6122
6122bis obsoletes 6122

So by following the chain, a reader of 5122 can discover that 6122bis is
the most up-to-date definition of the address format.

> If we need to do one or the other, I would also lean towards updating
> 5122 in 6122bis. Reving 5122 just for this seems like overkill.

Agreed.

Peter