Peter Saint-Andre wrote:
> Ralph Meijer wrote:
>> No, it seems to me that we have sailed beyond that
point and Dave made a
>> couple of nice observations that speak in favor of
having something like
>> <xmpp:pubsub.jabber.org?;node=mynode>. I am
just going ahead with using
>> such URIs to point to pubsub nodes in <link/>
elements in HTML for
>> auto-discovery purposes.
>
> OK. I think it would be good for us to document that in
XEP-0060 and
> also XEP-0030 (which are the two places where we define
nodes).
While chatting with Ralph over IM just now, I mentioned that
I think we
really need a little spec that clearly defines what a node
is, as used
in service discovery and publish-subscribe. So I'll that to
my to-do
list. Yay, yet another spec to write. 
Unfortunately we also have this thing called a node
identifier in a JID:
<node domain.tld>. This is confusing, and I think one of
them has got to
yield. The best Ralph and I could come up with is to change
the node
identifier portion of the JID to "localpart"
(along the lines of email),
then reserve the term "node" for these special
extra-JID thingies we use
in disco and pubsub. So we would have:
xmpp:localpartexample.org (e.g., a PEP-enabled account)
xmpp:localpartexample.org?;node=foo (a PEP node there)
Etc.
I realize that this breaks RFC 3920 naming (especially the
"nodeprep"
profile of stringprep). But I don't see a good way around
that, since I
think it's easier to change the nodeprep name than to change
the 'node'
attribute in disco and pubsub.
Peter
--
Peter Saint-Andre
https://stpeter.im/
|
