On Thu, Feb 14, 2008 at 11:22:09AM +0100, Ralph Meijer wrote: > > On Wed, 2008-02-13 at 11:57 -0700, Peter Saint-Andre wrote: > > Per recent list discussion, Ralph Meijer and I have provisionally > > adjusted XEP-0060 (Publish-Subscribe) to cover the use case of > > disassociating a node from a collection. > > [..] > > I see the attribute to refer to the node being dis/associated is still > called 'id', where I proposed 'node' for consistency with the rest of > the spec.
+1, will fix. > Furthermore, although you added <associate/> and <disassociate/> to the > schema definition of <collection>, the elements themselves don't have a > schema themselves. Oops, I will fix that. > The change for SHIM looks good to me, too. As I am looking into an HTTP > gateway to XMPP pubsub, I am wondering about the names of the headers > defined in this spec, though. They are very unlike other headers that > registered with the XMPP Registrar and headers used in other protocols > (HTTP, mail, news and MIME). > > RFC 3864 establishes the IANA registry of header fields and includes > advice on header field naming. As most of the headers in the SHIM > registry just point to their equivalent use in other protocols, I am > wondering if we should follow suit and move our registry there, while > adjusting the pubsub header fields to comply. I try not to bother the IANA more than necessary. However I will think about this. Is anyone actively using the headers yet? Will we break backwards-compatibility if we modify the header names now? /psa
